Identification of patients with persistent trophoblastic disease after complete hydatidiform mole by using a normal 24-hour urine hCG regression curve

• A 24-hour urine human chorionic gonadotropin regression curve can predict persistent trophoblastic disease after evacuation of a complete hydatidiform mole.
• At least one hCG-level exceeding the upper limit of p95 within 11 weeks could be a criterion of persistent disease.
• The curve can be used as a valuable visual tool to show patients their follow-up progress.

ObjectiveThe aim of this study was to establish a reference 24-hour urine human chorionic gonadotropin (hCG) regression curve in patients with complete hydatidiform mole (CHM) as diagnostic tool in the prediction of persistent trophoblastic disease (PTD).MethodsFrom 2004 to 2011, 312 cases suitable for this study were registered at the Hydatidiform Mole Registry of the Royal Women's Hospital Melbourne, Australia. hCG levels of 61 patients diagnosed as having PTD according to FIGO 2000 criteria were compared with the 95th-percentile (p95) of the normal regression curve derived from hCG levels of 251 cases of uneventful CHM.ResultsIn the test group of 61 patients PTD was diagnosed by FIGO 2000 criteria after a mean (± SD, min.–max.) of 7.6 (± 3.4, 3.0–16.7) weeks after evacuation of the mole while in the same group hCG values for the first time exceeded the upper limit of the 95th percentile significantly earlier after 4.5 (± 1.9, 2.0–9.9) weeks (P < 0.001).However, hCG levels of 14% of the cases of uneventful CHM at least once exceeded the upper limit of p95, showing that one single value above p95 is not accurate enough for the diagnosis of PTD.ConclusionsThe normal 24-hour urine hCG regression curve may be used as a tool in the follow-up of an individual case of CHM after evacuation. At least one hCG level exceeding the upper limit of p95 within 11 weeks after evacuation could be added to the current FIGO criteria, in order to diagnose PTD early, but the lack of it may also prevent unnecessary treatment.