Antral follicle count and anti-Müllerian hormone level as predictors of ovarian hyperstimulation syndrome in women with polycystic ovarian syndrome undergoing controlled ovarian stimulation

Ovarian hyperstimulation syndrome (OHSS) is a serious complication encountered in some women undergoing controlled ovarian stimulation (COS)⁽¹⁾. It is associated with an increased production of vasoactive substances such as angiotensin and vascular endothelial growth factor (VEGF), which leads to an increased capillary membrane permeability. The rise in permeability results in fluid accumulation in the third space and intravascular fluid depletion, thus making a woman susceptible to ascites, pleural effusion, and intravascular dehydration⁽²⁾. Mild OHSS occurs in 20–33% of women undergoing COS, moderate OHSS affects 3–6%, and only in 0.3% of all cases the syndrome is severe⁽³⁾. COS is commonly employed to increase the yield of ova available in order to enhance the success rate of artificial reproductive techniques.

The risk factors for OHSS include young age, polycystic ovarian syndrome (PCOS), decreased body mass index, and the presence of multiple follicles. The National Institute for Health and Care Excellence guideline on fertility predicts a likely high ovarian response to gonadotropin stimulation in an in vitro fertilisation (IVF) cycle and excessive ovarian response if any of the following is present: (i) a follicle-stimulating hormone (FSH) level of less than 4 IU/l, or (ii) an anti-Müllerian hormone (AMH) level greater than 25 pmol/l or (iii) an antral follicle count (AFC) greater than 16⁽⁴⁾. The problem with using these criteria – and specifically the AFC criterion – is that a woman is diagnosed with PCOS according to the Rotterdam criteria for PCOS⁽⁵⁾. As per the criteria; a woman is diagnosed when she has any two of the three features: 1) oligo- or anovulation, 2) clinical and/or biochemical signs of hyperandrogenism, or 3) polycystic ovaries; polycystic ovaries are defined as those found on the ultrasound to contain 12 or more follicles measuring 2 to 9 mm in diameter and/or have an increased volume of 10 mL or greater.

Only one ovary needs to meet these criteria for the definition of polycystic ovaries to be satisfied. Consequently, a woman with PCO may have 12 or more follicles on the initial scan and the criteria predictive of response may not fit in this particular scenario. This does not take away from the fact that PCOS is a high risk factor to begin with.

Now evidence is slowly emerging that these women may have a higher cut-off for the predictors as compared to the general population. A recent study showed that the cutoff for AMH in women with polycystic ovarian syndrome (PCOS) was higher. It has also been argued that the level of the anti-Müllerian hormone (AMH) is a better predictor of ovarian response than the antral follicle count (AFC)⁽⁶⁾. However, the data remains conflicting and studies continue to advocate both AFC and AMH as possible predictors of ovarian response⁽⁷⁾. Also, OHSS is more commonly seen in women with PCOS, but these women are also more likely to benefit from ovulation induction and have higher conception rates. We undertook this study to compare the rate of OHSS in women with PCOS undergoing ovulation induction and those without PCOS. We also set out to determine the cut-off of AMH and AFC for OHSS in women with PCOS and those without.

This was a prospective cohort study conducted at the Aziz Medical centre from the 1^st July 2017 until the 31^st December 2019. Women aged 20–39 years who were undergoing ovulation induction with clomiphene citrate were recruited. Women having polyps, thyroid dysfunctions, hyperprolactinaemia, premature ovarian failure, and ovulation induction with only gonadotrophins were excluded. Also excluded were women whose male partners were being treated for male factor infertility.

After obtaining informed consent, a brief history was taken from the women to ensure that they satisfied the study inclusion criteria. The women were explained the study protocol and follow-up. The patients undergoing induction were divided into those with PCOS and those without on the basis of the Rotterdam criteria. After that, they had a day 2 ultrasound to determine the antral follicle count. A Mindray DP-2200 scanner with a 5 to 7.5 MHz frequency endovaginal transducer was used for all the procedures. All scans were performed by the lead author who is a specialist in ultrasound and fertility, and has 10 years of professional experience in the field. For the estimation of serum AMH levels, all blood samples of selected subjects were drawn by venipuncture in serum separator tubes. Blood samples were taken for AMH levels on any day of the menstrual cycle. The serum AMH levels were determined by enzymelinked immunosorbent assay (ELISA), using human AMH Elisa kit (CDN-E 1350) at the reference lab.

The women were induced with clomiphene citrate 50 mg from menstrual cycle day 2 to day 6, for a total of 5 days. The women were followed up with a scan after every 48 hours to assess the size of follicles. Follicular monitoring was done, and ovulation was triggered with human chorionic gonadotrophin when the lead follicle reached a size of 18 mm. Sexual contact was encouraged, and the women were counselled about the signs of OHSS. They were requested to report back if they become pregnant, when a transvaginal scan was done to confirm a foetal heartbeat.

OHSS was classified according to the signs and symptoms tabulated in the RCOG guideline on OHSS⁽⁸⁾. The cases of OHSS were classified as mild, moderate or severe. A standard protocol was followed.

Clinical pregnancy was defined as foetal heartbeat present on a transvaginal scan after the women had a beta-hCG positive value in the serum.

For the estimation of sample size, we searched the literature and identified a study that compared OHSS in women with PCOS undergoing an IVF cycle and women with tubal factor infertility⁽⁹⁾ Using that study as a reference, and assuming similar proportions for the women with PCOS and those without, the sample size for the study came out to be n = 128 patients in each group. The sample size was calculated using the WHO software, where alpha = 5%, power of the test 1-beta = 80%, anticipated population proportion 1 = 6.3%, and anticipated population proportion 2 = 18.8%. We used the two group test (two-sided test) of equivalence in proportion to calculate the sample size (Sample Size Determination in Health Studies, Version 2.00, Copyright (c) 1996–98, World Health Organization). To compensate for protocol deviation, the sample size was inflated by 5%, so that minimum 135 women were needed in each study group

A proforma was used to collect the data. The demographic data of the women included age, height and weight. Reproductive history including the duration and type of infertility was also noted. The antral follicle count (AFC) on the initial scan, and the AMH level were also recorded in the proforma.

The primary outcome measure in this study was to assess to compare the rate of OHSS in women with and without PCOS. The secondary outcome measure was to assess the cut-off of AFC and AMH for OHSS in both groups.

All the participants provided their informed consent. In lieu of a formal ethics committee or formal institutional review board approval, the Declaration of Helsinki was followed. No subjects were harmed, confidentiality was maintained, and no patients were enrolled in the study without their formal informed consent.

Over the study period, a total of 711 women satisfied the inclusion criteria and were approached for participation. Of these, 12 women refused to participate and were excluded. Of these 699 women, a total of 282 had PCOS, and 417 were the controls. However, 6 cases and 4 controls did not comply with the follow-up and were excluded. Therefore, we included 689 women in the final analysis (Fig. 1).

Fig. 1.

Of these 689 women, 276 (40.1%) had PCOS, and 413 (59.9%) were the controls. The mean age of the study participants was 31.32 ± 3.82 years, and the mean BMI was 22.64 ± 1.71 kg/m². The mean duration of infertility was 3.19 ± 0.93 years, and the majority of the women (591, 85.8%) had primary infertility. The pregnancy rate of the study participants was 213 (30.1%), and OHSS occurred in 12.5% of them. Table 1 summarises the study population’s characteristics.

The cases and controls were not significantly different in terms of their age (p = 0.175) or duration of infertility (p = 0.245), but there was a significant difference in BMI (23.27 ± 1.94 vs. 22.21 ± 1.41, p <0.001) and type of infertility (p <0.001). OHSS occurred in 19.6% of the cases, and 7.7% of the controls (p <0.001). The mean antral follicle count among the cases was 14.32 ± 3.69 vs. 6.46 ± 2.47 (p <0.001), in the controls. The mean AMH was also significantly different in both groups (6.36 ± 1.10 ng/ml among the cases vs. 3.71 ± 0.45 in the controls, p <0.001) (Tab. 2).

A receiving operator curve (ROC) was used to predict hyper-response; Table 3 shows the area under the curve and 95%CI both for the cases and controls.

In women with PCOS, the sensitivity and specificity for the prediction of hyper-response were 94.4 % and 97.3 % for AFC and 92.6% and 93.7% for AMH, at the cut-off values of ≥18 and ≥6.425ng/ml, respectively. In the controls, the sensitivity and specificity for the prediction of hyperresponse were 93.8 % and 97.1% for AFC and 93.8% and 94.5% for AMH, at the cut-off values of ≥10 and ≥3.95 ng/ml, respectively.

The cut-off value for AFC and AMH in women with PCOS are 18 and 6.425 ng/ml, respectively, but the values for non-PCOS women are lower, 10 and 3.95 ng/ml, respectively. The findings show that group-specific values should be used to identify and counsel women undergoing COS.