Ustekinumab for Eczema? Why Psoriasis Success Did Not Establish Benefit in Two Trials
A medicine's success in psoriasis can sound promising to someone whose eczema remains difficult to control. Both conditions involve inflammation, but that shared description does not establish that the same treatment will work. Ustekinumab illustrates why the disease, comparator and clinical endpoint must stay attached to a claim.
A small study found biological changes without a significant primary difference
A phase 2 study published in the January 2017 issue randomized 33 adults with moderate-to-severe atopic dermatitis to ustekinumab (16) or placebo (17), followed by crossover. Background triamcinolone 0.025% cream was provided. The primary clinical endpoint was at least 50% improvement in objective SCORAD at week 16.
The between-group difference was not statistically significant: odds ratio 1.93, 95% confidence interval 0.30–15.33. That wide interval reflects substantial uncertainty. Molecular changes in skin samples supported biological activity, but did not establish a significant clinical advantage. Concurrent topical treatment and the small sample limit interpretation. A hypothesis about a different regimen remains a hypothesis, not a demonstrated patient benefit.
A separate Japanese trial also matters
The 2017 Japanese phase 2 trial randomized 79 adults aged 20–65 with severe or very severe atopic dermatitis: 24 received 45 mg ustekinumab, 28 received 90 mg and 27 received placebo. Injections were at weeks zero and four. The primary endpoint was percentage change in EASI at week 12.
Reported mean reductions were 38.2%, 39.8% and 37.5%, respectively. Neither active group established a significant advantage over placebo, and major secondary outcomes did not show significant differences. These within-group changes are not percentages of patients cured. The population and regimens constrain what the findings can answer; they do not test every possible schedule or every eczema subtype.
| Evidence item | What it supports | What it does not establish |
|---|---|---|
| Skin-sample molecular changes | A drug can affect biological pathways | A meaningful clinical advantage over the comparator |
| A lower score after treatment | Change occurred within a study group | The medicine caused more improvement than placebo |
| Two nonsignificant comparisons | These trials did not establish benefit | Exact equivalence or proof that no individual could respond |
Check the current indication rather than the medicine's reputation
The U.S. Stelara label updated in September 2026 lists plaque psoriasis, psoriatic arthritis, Crohn's disease and ulcerative colitis, with indication-specific age requirements. Atopic dermatitis is not listed. Its warnings include serious infections, tuberculosis evaluation, malignancy and serious hypersensitivity. Approved uses elsewhere in the body do not establish eczema efficacy or remove these risks.
If ustekinumab has been prescribed for another condition, this article is not a reason to stop it. Ask the relevant prescribers how a separate eczema diagnosis will be assessed and treated. Do not assume that one prescription automatically addresses both conditions.
Turn the evidence into better treatment questions
The AAD guideline discusses eczema-specific systemic options. Ask which evidence applies to your diagnosis, age, previous treatments and priorities. An injection's convenience or a familiar brand name cannot substitute for that discussion. Record which outcome matters most to you and how the clinician plans to measure it.
Our EASI explanation helps decode trial endpoints. The head-to-head treatment guide covers a different drug comparison. Our fezakinumab article examines another reason a promising biological argument or subgroup finding may fail to settle the overall clinical question.
Evidence and safety also need separate assessment. Our dupilumab eosinophil article examines a treatment-related laboratory question without using it as a substitute for clinical context.
Prepared by the Dr. Arsy editorial team. Sources checked November 9, 2026. Educational content does not provide an individual diagnosis or independent clinician review. Cited researchers and institutions do not endorse #003. The AI-created product image is an editorial illustration, not a patient photograph or clinical result.