Peptide LibraryOcular discomfort and corneal staining
Very low certaintyPatient-important outcome · 2 studies · 81 participants
The evidence conflicts
Two small trials have tested thymosin beta-4 eye drops against inactive drops in people with dry eye, and their results do not agree well. The larger trial missed its main goals, though some secondary measures looked better, while a very small trial reported improvements in discomfort and eye-surface damage. The certainty of this evidence is very low, so we cannot tell whether these drops really reduce dry eye discomfort or corneal staining.
What the research measured: In a 72-person placebo-controlled trial using the controlled adverse environment model, there was no significant difference on the primary endpoints (ocular discomfort score and inferior corneal staining at day 29). Discomfort during the chamber challenge was 27% lower with thymosin beta-4 (p=0.0244), and central (p=0.0075) and superior (p=0.021) corneal staining improved. In a 9-patient trial in severe dry eye (12 treated vs 6 vehicle eyes), day 56 OSDI discomfort was reduced by 35.1% (p=0.0141) and total corneal fluorescein staining by 59.1% (p=0.0108) versus vehicle. No confidence intervals were reported.
What remains unknown: Whether any benefit holds up in larger, fully reported trials with prespecified endpoints. Effects beyond about 8 weeks in this chronic condition. How results differ by dry eye type or severity, such as Sjogren's or graft-versus-host disease. Adequately reported safety data, since a lack of reported adverse events in small short trials does not establish safety.
Reviewers’ note: Starting at high for randomized evidence, downgrades for risk of bias, inconsistency, imprecision and publication bias exceed the available levels, so certainty is at the floor of very low.
How the certainty was judged
| Domain | Judgment | Why |
|---|---|---|
| Starting level | High (randomized evidence) | |
| Risk of bias | Very serious (rated down two levels) | Both trials are available only as abstracts, with no randomization, concealment or attrition details. In the larger trial the two prespecified primary endpoints were null, and the reported benefits come from selected secondary endpoints (judged high risk of bias) without multiplicity adjustment. The smaller trial is rated high risk of bias for all outcomes and singles out day 56 results from unspecified time points. |
| Inconsistency | Serious (rated down one level) | Results disagree. The 72-person trial found no difference on its primary discomfort and inferior corneal staining endpoints at day 29, but significant differences on some secondary measures. The 9-person trial reported significant improvements in discomfort and total corneal staining at day 56. |
| Indirectness | Serious (rated down one level) | One trial used an artificial controlled adverse environment challenge, and the other enrolled a small group with severe dry eye, including graft-versus-host disease. Both may not represent the general adult dry eye population, and the trials differ in outcome timing and in the staining regions measured. |
| Imprecision | Very serious (rated down two levels) | Only about 81 participants in total (72 randomized, plus 9 patients with 18 eyes). No confidence intervals are reported, and results are given mostly as p-values or percentage reductions. The smaller trial used eye-level comparisons in just 9 people. |
| Publication bias | Strongly suspected (rated down one level) | Both are small, abstract-only trials of a commercially coded product (RGN-259) with funding and conflicts not reported. Selective emphasis on favorable secondary endpoints and time points suggests reporting bias. |
Included studies
- Thymosin beta 4 ophthalmic solution for dry eye: a randomized, placebo-controlled, Phase II clinical trial conducted using the controlled adverse environment (CAE™) model. (Randomized controlled trial)
- Thymosin β4 significantly improves signs and symptoms of severe dry eye in a phase 2 randomized trial. (Randomized controlled trial, 9 participants)
Version 1 · published November 8, 2026 · evidence through November 8, 2026 · research snapshot 2026.1.1 · methodology ev-1 · GRADE (handbook, 2013) · Automated assessment by AI (claude-opus-5-5 and claude-sonnet-5-5), not reviewed by a person.