A biomarker's performance can depend on when it is measured. This study makes that issue visible by examining several visits after frozen-thawed embryo transfer. Primary study
What was observed
The analysis included 178 women: 21 with early miscarriage and 157 with viable singleton pregnancy. Serum kisspeptin and hCG were measured at 14, 21, and 28 days after transfer. Kisspeptin did not distinguish the groups at day 14. Later group differences appeared, but the reported discrimination was lower than hCG's at the corresponding visits. Primary report
Our appraisal: a later signal cannot repair an earlier prediction
When evaluating an early warning marker, the date on which information becomes available matters. A difference emerging later cannot be presented as a successful prediction at the first visit. The time remaining before an outcome or clinical diagnosis also affects practical usefulness.
A group difference and a useful individual prediction are separate claims. Overlapping distributions may produce a statistically detectable difference while leaving substantial uncertainty for an individual result. Conversely, an area-under-the-curve estimate does not specify the consequences of a particular clinical cutoff.
The hCG comparison is important because a new marker must be evaluated against information already available in the same setting. A lower standalone estimate does not rule out all possible added value in a combined model, but such added value must be tested rather than assumed.
The assisted-reproduction setting gives relatively structured timing. It does not guarantee the same performance in naturally conceived pregnancies or other patient groups. Assay differences and outcome-selection rules also require attention.
This report measured a naturally occurring biomarker; it did not test kisspeptin administration to prevent miscarriage. It should not support supplementation claims or individual pregnancy decisions. Full review should reconcile the reported design terminology, sample selection, visit-specific uncertainty, and any missing specimens before a detailed diagnostic interpretation.
Frequently asked questions
Was this a peptide-treatment trial?
No.
Did the day-14 measurement distinguish outcomes?
No.
Were later measurements different between groups?
Yes.
Did kisspeptin outperform hCG in the reported comparisons?
No.
Does a significant group difference guarantee accurate individual prediction?
No.
Were these pregnancies following frozen embryo transfer?
Yes.
Does lower measured kisspeptin establish a treatable deficiency?
No.