Predicting ovarian responses to the controlled ovarian hyperstimulation in elderly infertile women using clinical measurements and random forest regression

During the past decades, the number of elderly infertile women is obviously increasing in China, and more and more of them are likely to seek medical assisted reproductive technologies. As the in vitro fertilization/embryo transfer (IVF/ET) treatment presents special medical and psychological challenges to elderly infertile women, it is extremely helpful to perform the clinical evaluation and outcome prediction regarding IVF/ET outcomes. In this study, we retrospectively collected 12 clinical measurements in prior to the oocyte recovery for 689 elderly infertile patients (≥35 years of old), and used for predicting ovarian responses to the controlled ovarian hyperstimulation based on random forest regression models. Using different predictor sets and 10-fold cross validation approach, the Mean Square Error (±standard deviation) of prediction models varied from 7.56 ± 0.31 to 13.90 ± 0.37 in the training datasets, and the correlation coefficients between observed and predicted values ranged from 0.86 ± 0.02 to 0.72 ± 0.05 in the testing datasets. Among all clinical measurements involved in this study, the preovulatory follicle count (PFC), antral follicle count (AFC), and anti-Müllerian hormone (AMH) were revealed to be the most important features in prediction models. In conclusion, we successfully established the machine learning approach that could help the elderly infertile patients to better understand the most possible outcomes in subjecting to the controlled ovarian hyperstimulation.

Hormonal Effects in Reproductive Technology with Focus on Diminished Ovarian Reserve

Modern use of reproductive technologies has revolutionized the treatment of infertile couples. Strategies to improve ovarian function in cases of diminished ovarian reserve are perhaps the least understood area in this field and will be the chief focus of this chapter.

An AMH-based FSH dosing algorithm for OHSS risk reduction in first cycle antagonist protocol for IVF/ICSI

The study assessed the impact of an AMH algorithm for FSH dosing in 589 patients to maintain pregnancy rates while minimizing OHSS rates in 1st antagonist cycles for IVF. Patients with low AMH < 12 pmol/L (n = 203) had maximal stimulation with corifollitropin, patients with AMH 12-32 pmol/L (n = 256) had standard stimulation with 150 IU/day of rFSH and patients with AMH > 32 pmol/L (n = 130) had minimal stimulation with 112 IU/day of HP-hMG. The proportion of patients with targeted (5-14) number of oocytes at retrieval was: Low AMH 42%, intermediate AMH 76% and high AMH 67% (p < 0.001). Low responses (≤ 4 oocytes) was found in 55%, 16% and 26% (p < 0.001) in the low, intermediate and high AMH group, respectively. Excessive responses (≥15 oocytes) was found in 2.5%, 6.2% and 6.1% in the low, intermediate and high AMH groups, respectively. Despite the high proportion of low responses, the ongoing pregnancy rates in the high AMH group was 41% per started cycle. A total of 14 patients had OHSS preventive actions like agonist triggering (n = 12) and/or cryopreservation of all embryos (n = 4) and all avoided OHSS. Three (0.5%) patients were admitted to hospital with severe OHSS, and all occurred after hCG triggering and all cases were late OHSS in relation to pregnancy. All were in the high AMH group after aspiration of 10-15 follicles. The conclusion is that among high AMH patients, low dose HP-hMG will limit the mean number of oocytes, without compromising pregnancy rates. The OHSS risk will be low, but as long as transfer after hCG triggering is used OHSS will occur unless a cut-off for OHSS preventive actions as low as 10-15 follicles is used.

Is in utero exposure to maternal socioeconomic disadvantage related to offspring ovarian reserve in adulthood?

BACKGROUND

Because the ovarian follicle pool is established in utero, adverse exposures during this period may be especially impactful on the size and health of the initial follicle endowment, potentially shaping trajectories of ovarian follicle loss and the eventual onset of menopause. Building on a robust literature linking socioeconomic status (SES) and menopausal timing, the current study examined adverse prenatal exposures related to maternal SES, hypothesizing that greater maternal socioeconomic disadvantage would be associated with lower ovarian reserve in the adult offspring.

METHODS

In a healthy, community-based sub-sample (n = 350) of reproductive age participants in the OVA Study (2006-2011), prenatal maternal SES was examined in relation to two biomarkers of ovarian reserve, antimullerian hormone (AMH) and antral follicle count (AFC). Prenatal maternal SES was assessed indirectly using maternal addresses abstracted from participant birth certificates, geocoded, and linked to US Census-derived variables, including neighborhood-level characteristics: education (% of individuals with a HS diploma); poverty (% of families below the poverty line); unemployment (% of individuals > 16 years who are unemployed); and income (median family income).

RESULTS

In separate covariate-adjusted linear regression models (following the backward elimination of main effects with P > .10), greater maternal neighborhood education was related to higher ovarian reserve as marked by higher levels of offspring AMH (beta = .142, P < .001) and AFC (beta = .092, P < .10) with models accounting for 19.6% and 21.5% of the variance in AMH and AFC, respectively. In addition, greater maternal neighborhood poverty was related to lower ovarian reserve as marked by lower offspring AMH (beta = -.144, P < .01), with the model accounting for 19.5% of the variance in AMH.

CONCLUSIONS

Maternal socioeconomic disadvantage measured indirectly at the neighborhood level was associated with lower ovarian reserve among the adult offspring, independently of offspring SES and other potential confounding factors. This suggests SES-related adversity exposures may have a detrimental impact on the size or health of the initial follicle endowment, leading to accelerated follicle loss over time.

Does ovarian reserve affect outcomes in single ideal blastocyst transfers in women less than 40 years of age?

PURPOSE

There is much debate whether diminished ovarian reserve is purely a quantitative issue, or if quality as determined by pregnancy potential of the ensuing oocytes is also affected. The purpose of this study was to determine whether diminished ovarian reserve, as established by one of three ways described below, affects pregnancy outcomes of women under 40 years old undergoing a single ideal blastocyst transfer.

MATERIALS AND METHODS

This was a retrospective cohort study, including 507 women undergoing an ideal quality single embryo transfer between August 2010 and March 2014. Logistic regression was used to control for age, duration of infertility, parity, body mass index, and smoking status. For analysis, women were stratified for: antral follicle counts (≤ 5 vs. > 5), basal serum FSH levels (< 13 vs. ≥ 13 IU/L), and quartile of total FSH dose required for stimulation.

RESULTS

In stratifying women by antral follicle count (AFC) ≤ 5 vs. > 5, the pregnancy rate (40 vs. 53%, p = 0.04), clinical pregnancy rate (29 vs. 46%, p = 0.02), and live birth rate (13 vs. 43%, p = 0.001) were superior with AFC > 5. Using FSH levels (< 13 vs. ≥ 13 IU/L), the pregnancy rate (50 vs. 31%, p = 0.27), clinical pregnancy rate (40 vs. 13%, p = 0.45), and live birth rate (38 vs. 13%, p = 0.48) were similar. Examining quartiles of FSH stimulation, the pregnancy rates were similar (from lowest to highest: 45, 52, 54, 41%, p = 0.13); however, clinical pregnancy rate (36, 43, 47, 25%, p = 0.003) and live birth rate (32, 38, 44, 20%, p = 0.005) were superior in lower quartiles vs. the highest quartile.

CONCLUSION

Ovarian reserve may affect embryo pregnancy potential and outcomes when measured by AFC and exogenous stimulation but not by basal FSH levels.