Objective: To analyse prognostic factors in complete hydatidiform moles usi
ng multiple logistic regression analysis. Methods: Evaluation of host and t
umour related parameters including (a) gestational age, patient age, parity
, molar phenotype, grade of proliferation of the tumour and cytological aty
pia, (b) expression of beta-HCG, EGF, EGFR, TGF-alpha, TGF-beta, IL1-alpha,
IL1-beta by immunohistochemistry, (c) serial monitoring of serum beta-HCG
levels by ELISA, and (d) lectin binding using jack fruit lectin histochemis
try as indices for persisting trophoblastic disease (PTD). Results: Serum b
eta-HCG levels at 4 weeks, cellular atypia, lectin binding, expression of T
GF-alpha and IL1-beta showed highly significant correlation with persistenc
e of the tumour (P<0.001). The sensitivity and specificity at 4 weeks in co
mbination with cytological atypia to identify spontaneously regressing lesi
ons was 100% and those requiring chemotherapeutic intervention was 80%. Con
clusion: The concentration of serum beta-HCG 4 weeks post evacuation (<300
mIU/ml) combined with cytological abnormalities could identify nearly 100%
of the spontaneously regressing lesions (low risk) and 80% of those needing
chemotherapeutic intervention (high risk), thereby suggesting that patient
s who have a serum P-HCG at 4 weeks of evacuation <300 mIU/ml with no cytol
ogical atypia of the trophoblasts need only be followed up at long interval
s, while those having a serum beta-HCG at 4 weeks of evacuation >300 mIU/ml
accompanied with cytological atypia of the trophoblasts should be closely
followed up. (C) 1999 Elsevier Science ireland Ltd. All rights reserved.