Serum progesterone for the exclusion of early pregnancy in women at risk for recurrent gestational trophoblastic neoplasia.

OBJECTIVE

To evaluate the utility of the serum progesterone level for discriminating pregnancy from gestational trophoblastic neoplasia.

METHODS

Serum progesterone levels were measured in 61 women with histories of trophoblastic disease who developed a re-elevation in hCG during surveillance and underwent a work-up to differentiate pregnancy from gestational trophoblastic neoplasia. Progesterone levels were analyzed in the context of diagnostic outcome (pregnancy versus gestational trophoblastic neoplasia) to identify optimal threshold levels of progesterone to be used for classifying outcome.

RESULTS

Of the 61 women, 37 proved to be pregnant and 24 had gestational trophoblastic neoplasia. Progesterone less than 2.5 ng/mL was predictive of trophoblastic malignancy, with a sensitivity of 83% (20 of 24 subjects were classified correctly as having gestational trophoblastic neoplasia) and a specificity of 95% (35 of 37 patients with progesterone levels at or above 2.5 ng/mL were correctly classified as pregnant). Progesterone of at least 10 ng/mL was associated with viable pregnancy in 97% of the cases. Furthermore, the progesterone level predicted outcome regardless of the serum hCG value.

CONCLUSIONS

The serum progesterone level is useful for discriminating early pregnancy from gestational trophoblastic neoplasia.