Background:  The major concern in gestational trophoblastic disease is management of persistent disease and malignant sequelae. However, prediction of response to treatment is difficult and methods used controversial.
Aim and methods:  To evaluate the usefulness of clinical presentation, methods of diagnosis and categorisation of risk in determining clinical outcomes, by analysis of a database of 705 registered patients collected over 30 years.
Results:  From the database, there were 97 patients who developed persistent disease and malignant sequelae on the basis of defined criteria – 80.4% had molar pregnancy and 19.6% non‐molar pregnancy. Vaginal bleeding was not a common presentation; 59.8% had no clinical symptoms. According to protocol, monitoring by serial human chorion gonadotrophin (HCG) levels followed by imaging screen was used in all patients; histology was also available in 41.2% from hysterectomy and curettage specimens. There were 16 of 76 patients with persisting disease who had metastases (21.1%), and 2 of 20 patients with choriocarcinoma who had an antecedent molar pregnancy (10.0%). Based on five risk factors, 25 patients were categorised as ‘high risk’ and assigned to receive multi‐drug chemotherapy. There were two deaths (2.1% for all malignant sequelae); both were from molar pregnancies. One patient failed to respond and the other suffered a complication of intensive chemotherapy.
Conclusion:  Serial HCG levels remain the best monitor to determine therapeutic response. Categorisation of ‘high risk’ by five factors is useful in treatment. Albeit a small series, clinical outcome is favourable with a five‐year survival of 89.7%.