The thymosin alpha-1 liver-failure report illustrates why a favorable survival statement needs its endpoint and analysis attached. The abstract does not describe every survival analysis as favorable. Primary trial
What the report says
The open-label randomized study enrolled patients with hepatitis-B-related acute-on-chronic liver failure and compared a thymosin alpha-1 strategy with standard medical therapy. It reports a higher 90-day liver-transplantation-free survival estimate in the active group, but no significant survival difference using its competing-risk analysis. New infection and hepatic encephalopathy findings were also reported. The abstract gives 120 enrolled participants but group counts totaling 114, requiring participant-flow reconciliation. Primary trial
Our appraisal: do not choose only the favorable analysis
Transplantation, death, and remaining alive without transplantation are not interchangeable events. Before interpreting an estimate, identify exactly which event ends follow-up and how the other events are handled. Different analysis choices can answer different clinical questions.
For a hypothetical illustration, an intervention could change how frequently patients undergo transplantation without producing the same change in overall mortality. A combined endpoint might then move differently from one of its components. This example is not a claim about what caused the results in this trial; it explains why endpoint definitions matter.
Questions that remain open in this record
The complete report should clarify the enrolled, randomized, treated, and analyzed populations and account for the differing counts in the abstract. That mismatch should not be silently repaired by assuming which participants were excluded or why.
We would also inspect the prespecified primary analysis, definitions of infection and encephalopathy, missing-data handling, concurrent care, and the implications of open-label assessment. Subgroup explanations should be checked for prespecification and credibility rather than treated as established treatment-selection rules.
This is a historical trial appraisal. It is not guidance for managing liver failure, a current treatment recommendation, or evidence that any separate research material is appropriate for human administration.
Frequently asked questions
Did every survival analysis show a significant difference?
No. The abstract reports different conclusions for different analyses.
Is transplantation-free survival identical to all-cause survival?
No. The endpoint definition includes an additional event boundary.
Can the reported sample counts be added without checking flow?
No. The abstract's enrollment and group totals need reconciliation.
Does that discrepancy itself prove misconduct?
No. It identifies a reporting question that requires investigation.
Does open-label mean unrandomized?
No. Masking and random allocation are separate design features.
Can a proposed subgroup benefit guide self-treatment?
No. Exploratory interpretation is not an individualized treatment rule.
What is the source limit here?
The primary indexed abstract and record. Full participant-flow and analysis verification are explicit release requirements.