Probiotics for Eczema: Which Trial Results Apply to Which People?

Probiotics for eczema have mixed treatment results, not a dependable promise of clearer skin. Selected trials found benefits from particular oral formulations, while an infant trial found no benefit and a Cochrane review found little or no improvement in patient-rated symptoms. The evidence concerns specific organisms and populations; it does not establish that any probiotic supplement, yogurt, or fermented food treats eczema.

Cochrane: probiotics for treating eczema (2018) · Navarro-Lopez et al.: pediatric randomized trial · Gruber et al.: infant randomized trial · NIH ODS: probiotic definitions and product identification · Sources checked September 21, 2026.

What exactly is being tested?

NIH identifies probiotics by genus, species, and strain. A strain designation is part of the organism’s identity, not decorative label text. Prebiotics feed microorganisms; postbiotics contain nonliving microorganisms or their components. Those categories are not interchangeable. This page examines selected oral treatment studies in people who already had atopic dermatitis, rather than claiming to be an exhaustive review.

NIH ODS: probiotic definitions and product identification · Sources checked September 21, 2026.

Older children: a positive result for a named mixture

A randomized trial published in JAMA Dermatology in 2018 enrolled 50 children aged 4–17 with moderate atopic dermatitis. The intervention was a daily capsule containing a total of 1 billion colony-forming units (CFU): Bifidobacterium lactis CECT 8145, B. longum CECT 7347, and Lactobacillus casei CECT 9104. The comparison was a maltodextrin placebo over 12 weeks.

The reported reduction in the SCORAD severity index was 19.2 points greater with the mixture than with placebo. Topical steroid use was also recorded. The publication disclosed that several authors worked for Biopolis. These are findings for this formulation and study population, not instructions to give a child probiotics.

Navarro-Lopez et al.: pediatric randomized trial · Sources checked September 21, 2026.

Infants: the same treatment duration did not produce a benefit

A 2007 placebo-controlled trial studied infants aged 3–12 months with mild-to-moderate atopic dermatitis. It tested Lactobacillus rhamnosus GG at 5 billion CFU twice daily for 12 weeks; 54 infants received LGG and 48 received placebo. Symptoms improved over time, but the study found no statistically significant treatment-group difference in its combined symptom measure or rescue-medication outcomes.

Gruber et al.: infant randomized trial · Sources checked September 21, 2026.

Adults: a different organism and a small trial

A randomized trial in 30 adults with mild-to-moderate atopic dermatitis tested microencapsulated Lactobacillus plantarum IS-10506 at 20 billion CFU per day against a skim milk-Avicel placebo. SCORAD was significantly lower in the probiotic group at week eight. The abstract reports the direction and statistical significance, but does not give a numeric between-group SCORAD difference; none is supplied here.

Prakoeswa et al.: adult randomized trial · Sources checked September 21, 2026.

Our comparison: why these trials cannot rank products

The infant study and the older-child study both lasted 12 weeks, yet their results differed. Duration alone therefore does not reconcile them. The participants, baseline disease severity, organisms, and outcome definitions also changed. These studies cannot tell us which of those differences explains the conflicting findings.

Nor can the adult result be used as the next step on a dose ladder. It tested another strain in another age group, not a higher dose of the pediatric mixture. There was no shared treatment arm connecting the studies. Calling the largest CFU amount the strongest eczema treatment would invent a comparison the investigators did not perform.

A mixture result also leaves an attribution problem: the trial compared the whole combination with placebo. It did not isolate the contribution of each organism. Replacing the combination with one component would change the intervention. Keeping the full strain names is what makes this limitation visible.

Navarro-Lopez et al.: pediatric randomized trial · Gruber et al.: infant randomized trial · Prakoeswa et al.: adult randomized trial · Sources checked September 21, 2026.

Why a positive trial and a cautious review can coexist

The 2018 Cochrane review included 39 trials and 2,599 participants, using searches through January 2017. Its conclusion was that probiotics probably made little or no difference in patient-rated eczema symptoms. Investigator-rated severity showed a small effect of uncertain clinical importance. This is a dated evidence synthesis, not a claim that no later research exists.

Cochrane: probiotics for treating eczema (2018) · Sources checked September 21, 2026.

Our comparison: measure the claim before comparing the number

A change in a composite severity score and a change noticed by a patient are related questions, but they are not identical outcomes. The positive pediatric trial above emphasizes SCORAD and steroid-use days; the review separately evaluates patient-rated symptoms. Treating these as one interchangeable success measure conceals what was actually assessed.

The adult abstract supplies a statistically significant difference without the point estimate needed here to compare effect sizes. That missing number stays missing. Assigning it the pediatric trial’s improvement would be an invented result, even though both studies used the name SCORAD.

The review’s search cutoff also matters. Its conclusion describes the studies available to that analysis, while a later publication belongs to a later evidence record. A new positive study does not automatically erase the earlier null results, and an older pooled estimate cannot answer every question about a newly tested formulation.

Cochrane: probiotics for treating eczema (2018) · Navarro-Lopez et al.: pediatric randomized trial · Prakoeswa et al.: adult randomized trial · Sources checked September 21, 2026.

Our comparison: three claims that need separate evidence

Prevention asks whether eczema develops; treatment asks whether existing eczema improves. A result about future disease onset cannot provide a treatment response rate for a person who already has it. Likewise, applying bacteria to skin is a different intervention from swallowing a capsule. Neither comparison supports merging all microbiome products into a single success claim.

Fermented food introduces another gap. A food may contain living organisms, but the label does not necessarily identify the trial strain or its delivered quantity. The studies above tested specified preparations against specified controls. Turning their capsule results into a yogurt serving size would require evidence that these studies do not supply.

NIH ODS: probiotic definitions and product identification · Navarro-Lopez et al.: pediatric randomized trial · Gruber et al.: infant randomized trial · Prakoeswa et al.: adult randomized trial · Sources checked September 21, 2026.

Safety is part of the evidence, too

NCCIH reports limited long-term safety data and greater concern in people with underlying health conditions. Severe or fatal infections have been reported in premature infants receiving probiotics. Its guidance is to discuss supplements with a healthcare professional and not replace established treatment with unproven products. Trial doses on this page describe research, not a personal dosing plan.

NCCIH: probiotic safety and limits · Sources checked September 21, 2026.

The unanswered question on a supplement label

For a product making an eczema claim, the missing bridge is often an exact match: the named preparation, the population studied, and the outcome demonstrated. A familiar bacterial species or a large CFU number does not fill that bridge. These selected trials leave no defensible all-ages ranking of the best probiotic for eczema.

NIH ODS: probiotic definitions and product identification · Navarro-Lopez et al.: pediatric randomized trial · Gruber et al.: infant randomized trial · Prakoeswa et al.: adult randomized trial · Sources checked September 21, 2026.

Reviewed September 21, 2026. Educational analysis; not individualized medical advice.