A trial sponsored by Caspian Therapeutics found that its drug reduced a laboratory biomarker by 21% relative to placebo. The trial was too short to measure the clinical outcomes, hospitalization and functional decline, that matter most to patients, and mild adverse events were more common in the treatment group. Biomarker improvement may justify a larger trial, but it cannot by itself establish that patients benefit. A surrogate measure is useful only to the extent that changes in it reliably predict changes in meaningful outcomes.