SERUM ESTRADIOL LEVEL AND OOCYTE NUMBER IN PREDICTING SEVERE OVARIAN HYPERSTIMULATION SYNDROME

Ovarian hyperstimulation syndrome (OHSS) is a relatively common and po tentially life-threatening complication of ovarian stimulation, the pa thogenesis of which remains unclear. To clarify the predictive values of serum estradiol levels and oocyte number in severe OHSS, and to inv estigate the impact of high serum estradiol levels on pregnancy outcom e, we retrospectively analyzed clinical data from 431 cycles of ovaria n stimulation for assisted reproduction performed for 1993 through 199 5. Receiver operating characteristic plot were used to estimate the pr edictive power of the measured variable. The overall frequency of seve re OHSS was 5.5%. Using a serum estradiol level of 3,600 pg/mL as the minimum cut-off value, the sensitivity was 58%, with a specificity of 92%, a positive predictive value of 29%, and a negative predictive val ue of 97%. The predictive power was similar when a cut-off point of 20 oocytes retrieved was used. The two criteria together gave a sensitiv ity of 33%, a specificity of 92%, a positive predictive value of 40%, and a negative predictive value of 98%. One of seven oocyte donors dev eloped severe OHSS. The pregnancy rate was higher in patients with sev ere OHSS than in patients who did not develop this syndrome (73.9% vs 32.5%) but the pregnancy outcomes were not significantly different. We conclude that elevated estradiol concentrations and oocyte number app ear to be helpful in predicting severe OHSS, but neither parameter by itself is predictive. This syndrome is rare in the absence of luteal h CG support, either exogenous or pregnancy-derived; when it occurs, the re are usually extremely high preovulatory estradiol concentrations an d numerous oocytes retrieved. High serum estradiol levels are unlikely to have adverse effects on pregnancy outcome in patients with severe OHSS.