A promising pregnancy marker must add useful information at the time a decision is needed. This IVF/ICSI study tested repeated kisspeptin measurements during very early pregnancy. Primary study
What was studied
Seventy-three participants were classified as having biochemical pregnancies, early miscarriages, or viable pregnancies. Blood was sampled at the pregnancy test and four days later. Initial kisspeptin was lower in biochemical pregnancies, but it did not significantly distinguish early miscarriage from viable pregnancy. Serial kisspeptin measurements did not perform better than beta-hCG. Primary report
Our appraisal: adding a measurement is not necessarily adding information
The appropriate benchmark for a new biomarker is not simply whether it differs between any two groups. It is whether it improves the particular decision facing a patient and clinician, beyond established information.
A difference between biochemical and viable pregnancies does not automatically solve the harder discrimination between a pregnancy that will miscarry and one that will continue. Pooling these outcomes into a broad unfavorable category can obscure where a test succeeds and where it fails.
Repeated sampling introduces another distinction. A second result can improve characterization of a trajectory without improving prediction enough to justify additional testing. The comparison should concern the performance of the complete proposed testing strategy, including its timing, rather than isolated significance tests at individual visits.
The result should also stay attached to the studied clinical window. It does not rule out different behavior later in gestation, and later-pregnancy findings cannot retrospectively turn this early test into a useful one. Neither direction of extrapolation is warranted.
For readers comparing pregnancy biomarker papers, the central questions are when samples were taken, which outcomes were distinguished, and what existing test served as the comparator. This report provides an important negative answer to a specific incremental-value question, not evidence for administering kisspeptin during pregnancy.
Frequently asked questions
Was this an intervention trial?
No, it evaluated measurements.
Were samples repeated?
Yes, at two early time points.
What was the established comparator?
Beta-hCG.
Did serial kisspeptin outperform it?
No.
Were miscarriage and viable pregnancy clearly separated?
Not significantly by the initial kisspeptin measurement.
Does this settle every gestational window?
No. Timing affects the question being tested.
What is the main lesson?
Biomarker association is not the same as added clinical value.