Do ovarian reserve tests predict miscarriage in women undergoing assisted reproduction treatment?

To evaluate the role of ovarian reserve tests for the prediction of miscarriage among pregnancies resulting from assisted reproduction treatment.

Cohort study in a large IVF centre. Three hundred and fourteen women aged <43 years with basal FSH <12 IU/L who conceived following their first cycle of IVF/ICSI were included. Sixty-seven participants experienced miscarriage while the remaining 247 had ongoing pregnancies. Intervention included transvaginal scan and venepuncture on day 3 of the menstrual cycle immediately prior to treatment. Ovarian reserve parameters assessed were antral follicle count (AFC), FSH, basal oestradiol and oestradiol response. The main outcome measure was miscarriage rate (pregnancy loss at <12 weeks).

While the baseline clinical characteristics including age and duration of subfertility were comparable, BMI was significantly higher in women who experienced miscarriage compared to those having ongoing pregnancies. Among the ovarian reserve markers studied, only the AFC was significantly different between the miscarriage and pregnant groups. Logistic regression analysis revealed that only these two markers (BMI and AFC) were the significant predictors of clinical miscarriage. However, the discriminative ability of both BMI and AFC for the prediction of miscarriage was low as indicated by AUCs of 0.617 and 0.588 respectively on ROC curve analysis.

While AFC and BMI are significantly predictive of first trimester miscarriage, the ability of these markers individually or in combination to discriminate women who experience miscarriage from those who continue the pregnancy beyond 12 weeks of gestation is poor.