Surveillance for gestational trophoblastic neoplasia following molar pregnancy: a cost-effectiveness analysis

Historically, published guidelines for care after molar pregnancy recommended monitoring human chorionic gonadotropin levels for the development of gestational trophoblastic neoplasia until normal and then for 6 months after the first normal human chorionic gonadotropin. However, there are little data underlying such recommendations, and recent evidence has demonstrated that gestational trophoblastic neoplasia diagnosis after human chorionic gonadotropin normalization is rare.

We sought to estimate the cost-effectiveness of alternative strategies for surveillance for gestational trophoblastic neoplasia after human chorionic gonadotropin normalization after complete and partial molar pregnancy.

A Markov-based cost-effectiveness model, using monthly cycles and terminating after 36 months/cycles, was constructed to compare alternative strategies for asymptomatic human chorionic gonadotropin surveillance after the first normal (none; monthly testing for 1, 3, 6, and 12 months; or every 3-month testing for 3, 6, and 12 months) for both complete and partial molar pregnancy. The risk of reduced surveillance was modeled by increasing the probability of high-risk disease at diagnosis. Probabilities, costs, and utilities were estimated from peer-reviewed literature, with all cost data applicable to the United States and adjusted to 2020 US dollars. The primary outcome was cost per quality-adjusted life year ($/quality-adjusted life year) with a $100,000/quality-adjusted life year willingness-to-pay threshold.

Under base-case assumptions, we found no further surveillance after the first normal human chorionic gonadotropin to be the dominant strategy from both the healthcare system and societal perspectives, for both complete and partial molar pregnancy. After complete mole, this strategy had the lowest average cost (healthcare system, $144 vs maximum $283; societal, $152 vs maximum $443) and highest effectiveness (2.711 vs minimum 2.682 quality-adjusted life years). This strategy led to a slightly higher rate of death from gestational trophoblastic neoplasia (0.013% vs minimum 0.009%), although with high costs per gestational trophoblastic neoplasia death avoided (range, $214,000 to >$4 million). Societal perspective costs of lost wages had a greater impact on frequent surveillance costs than rare gestational trophoblastic neoplasia treatment costs, and no further surveillance was more favorable from this perspective in otherwise identical analyses. No further surveillance remained dominant or preferred with incremental cost-effectiveness ratio of <$100,000 in all analyses for partial mole, and most sensitivity analyses for complete mole. Under the assumption of no disutility from surveillance, surveillance strategies were more effective (by quality-adjusted life year) than no further surveillance, and a single human chorionic gonadotropin test at 3 months was found to be cost-effective after complete mole with incremental cost-effectiveness ratio of $53,261 from the healthcare perspective, but not from the societal perspective (incremental cost-effectiveness ratio, $288,783).

Largely owing to the rare incidence of gestational trophoblastic neoplasia after human chorionic gonadotropin normalization after molar pregnancy, prolonged surveillance is not cost-effective under most assumptions. It would be reasonable to reduce, and potentially eliminate, current recommendations for surveillance after human chorionic gonadotropin normalization after molar pregnancy, particularly among partial moles. With any reduction in surveillance, patients should be counseled on symptoms of gestational trophoblastic neoplasia and established in routine gynecologic care.