Delaying the oocyte maturation trigger by one day leads to a higher metaphase II oocyte yield in IVF/ICSI: a randomised controlled trial

Various ultrasound and hormonal criteria have been used to determine the moment to trigger oocyte maturation in IVF/ICSI cycles. Historically, the moment for triggering oocyte maturation has been based on follicle diameters that were measured using ultrasound and levels of serum estradiol [1, 2]. A Cochrane review [3] stated that the use of sonographic criteria alone might be sufficient but that the simultaneous determination of serum estradiol is still recommended as long as large randomised controlled trials have not shown that the incidence of ovarian hyperstimulation syndrome is equal in both groups. Combined monitoring was recommended as "a precautionary good practice point". These findings were mainly gathered in agonist protocols.

The decision to advance trigger oocyte maturation by 24 hours did not seem to have a positive effect on the probability of pregnancy in an antagonist protocol [4].

Irrespective of the use of an agonist or an antagonist for suppression, we tested whether the protocol of Garcia-Velasco et al. [5], in which oocyte maturation is triggered as soon as 3 follicles reach diameters of 18 mm, could be adapted further. These authors used it to compare ovarian steroid production when either an agonist or an antagonist was used. We also applied a uniform protocol for monitoring and hypothesised that delaying the administration of the trigger for 24 hours would result in a higher yield of mature oocytes, which served as our primary outcome variable. To evaluate whether this modification had any consequences on pregnancy rates or pregnancy outcomes, these variables were further evaluated as secondary outcomes. This randomised controlled trial was performed in patients with normal serum progesterone levels.

It is known that high progesterone levels (>1.5 ng/ml) have a deleterious effect on the endometrium and, thus, on pregnancy rates [6]. When progesterone levels are slightly elevated (>1 ng/ml), it can be difficult to decide whether to continue the stimulation procedure for one more day. A randomised controlled trial was carried out on this group of patients to evaluate the number of mature oocytes at retrieval as the primary outcome variable. Pregnancy rates and outcomes were important secondary variables.

This study was approved by the Ethical Committee of the Ghent University Hospital (B67020108975) and as a clinical trial internationally (NCT01980563 at ClinicalTrials.gov). It was part of a larger prospective trial in a single university hospital (Gent, Belgium) that compared cycle monitoring for IVF/ICSI in two parallel control groups: those with ultrasound monitoring versus those with combined monitoring (ultrasound plus hormonal monitoring). Between January 31, 2011, and December 31, 2011, 262 consecutive patients from the latter group were eligible for the present trial.

Inclusion criteria were the presence of both ovaries and being a female less than 45 years of age. Patients with ovarian cysts at the start of the ovarian stimulation procedure were excluded. Ovarian reserve was determined by measuring anti-Müllerian hormone (AMH) before starting treatment (Immunotech, Beckman Coulter Company, Brea, CA, USA).

Various protocols for controlled ovarian hyperstimulation were applied. Either recombinant FSH (Gonal F®, Merck Serono, Geneva, Switzerland) or urinary FSH (Menopur®, Ferring Pharmaceuticals, Saint-Prex, Switzerland) was used with daily doses between 150 and 300 U, dependent on age, anti-Müllerian hormone (AMH) levels and previous response, if applicable. In the agonist group, 0.1 mg triptorelin (Decapeptyl®, Ipsen, Paris, France) was administered subcutaneously for 7 days starting on cycle day 1, and gonadotrophins were started on cycle day 3. In the antagonist group, a fixed protocol was used: gonadotrophins were started on cycle day 3, and 0.25 mg cetrorelix (Cetrotide®, Merck Serono, Geneva, Switzerland) was injected subcutaneously as a daily dose from the 6th day of stimulation until the day of oocyte maturation triggering. After 1 week of stimulation with gonadotrophins, a first ultrasound monitoring session was planned. Serum levels of estradiol, LH and progesterone were determined simultaneously. All samples were analysed with ECLIA (Modular E170, Roche, Vilvoorde, Belgium). The inter- and intra-assay coefficients of variability for the progesterone assay were 3.46–6.71% and 1.1–7.0%, respectively. The cut-off for the sensitivity of the test (minimal detectable level) was 0.15 ng/ml. Depending on the findings, patients were scheduled for additional monitoring every 1 or 2 days. As soon as three follicles reached a diameter of at least 18 mm, patients were divided into two groups: those with serum progesterone > 1 ng/ml and those with a low progesterone level, defined as ≤ 1 ng/ml. The results of the individual monitoring of the patients were centralised and discussed at a daily staff meeting. All 6 staff members who were performing monitoring enrolled patients equally.

When the serum progesterone level was > 1 ng/ml, patients were randomised (single blinded). They received 5000 U of human chorionic gonadotrophin (hCG) (Pregnyl®, Merck Sharp & Dohme, NJ, USA) on the same day (high progesterone early [HPE] group) or 24 hours later (high progesterone late [HPL] group). A computer-generated list was used, and allocation concealment was performed by a secretary over the telephone. This procedure was supervised by a single observer (FV). No further monitoring was planned the day after randomisation into the HPL group. When more than 3 oocytes were expected, oocytes were retrieved with a single lumen needle. Flushing through a double lumen needle was performed in all other cases. All laboratory procedures were carried out as previously described [7]. In cases of ICSI treatment, the number of mature oocytes was dependent on their morphological appearance after denudation. In IVF cycles, all oocytes that were inseminated were classified as mature, as is acceptable convention per the literature. A maximum of 2 embryos were transferred 3 days after oocyte retrieval. Luteal support consisted of 600 mg micronised progesterone (Utrogestan®, Besins Healthcare, Bangkok, Thailand) that was administered vaginally in three daily doses, starting after oocyte collection, and continued until 2 weeks after transfer if not pregnant or until a clinical pregnancy was confirmed by ultrasound.

For patients with a low progesterone level (<1 ng/ml), a different approach was used. When > 50% of the follicles were larger than 18 mm, 5000 U of hCG was injected on the same day. Only follicles of at least 10 mm were counted to obtain this ratio. If the number of follicles with a diameter of at least 18 mm was between 30 and 50% of the total number counted, the patient was randomised. They received 5000 U of hCG on the same day (low progesterone early [LPE] group) or 24 hours later (low progesterone late [LPL] group), and no extra monitoring procedure was performed the day after randomisation in the LPL group. Allocation was performed as previously described. Further treatment was completed as described above for patients with a progesterone level > 1 ng/ml.

The results of the trial were processed anonymously by a single observer (FV). The number of metaphase II oocytes (MII) was our primary outcome variable. Secondary variables that demonstrate the oocyte yield were the number of oocytes retrieved, the number of fertilised oocytes (2PN) and the number of good quality embryos (GQE) on day 3. Other secondary outcome variables were defined according to the literature [8]: pregnancy rates (PR), clinical pregnancy rates (CPR), ongoing pregnancy rates (OPR) and live birth rates (LBR) as expressed per cycle; clinical implantation ratios (CIR/E), ongoing implantation ratios (OIR/E) and live birth ratios (LBR/E) were calculated for each individual embryo that was transferred.

The sample size calculation for the group with progesterone levels ≤ 1 ng/ml was based on a mean yield of 6 mature oocytes (SD = 3) in the LPE group versus 11 (SD = 6) in the LPL group [4, 9], resulting in a required sample size of 15 in each group (Welch’s t-test, 5% significance level, 80% power). For patients with high progesterone levels (>1 ng/ml), we found no comparable data in the literature as a reference. We therefore decided to include patients in the HPE and HPL groups concomitantly with the LPE and LPL groups.

The descriptive analyses in Table 1 were based on Fisher's exact tests for proportions and Student’s t-tests for continuous outcomes; when the data were skewed or contained outliers, the non-parametric Mann–Whitney U-test was used. To account for correlations between measurements for women with repeated cycles, all further analyses were based on linear and logistic marginal regression models that were fitted using generalised estimating equations with exchangeable working correlations. Although adjustment for baseline covariates was not required in view of the randomised study design, adjustments for age, PCOS and female pathology were used to improve precision in some of the linear models. All tests were performed at the 5% significance level. Statistical analyses were performed using SPSS, version 21, and R Studio, version 0.97.320.

Subsequent to ending the trial, additional evidence came to suggest that serum progesterone levels > 1.5 ng/ml at the moment of triggering oocyte maturation might lower pregnancy rates. Therefore, we performed an additional comparison of 2 subgroups of patients: those with low progesterone levels (<1.0 ng/ml; group A) and those with highly elevated levels (>1.5 ng/ml; group B). Fisher's exact test and Student's t-test were again used as described above.

Seventy-two patients with at least three follicles ≥ 18 mm had serum progesterone levels > 1 ng/ml. They were randomised into 2 groups (HPE and HPL) of 36 individuals each. In the HPE group, the oocyte maturation trigger was administered on the same day. In the HPL group, hCG was injected 24 hours later. In the remaining cases with low serum progesterone levels (<1 ng/ml), 59 patients had 30 to 50% of their follicles measuring equal to or greater than 18 mm. After randomisation, 28 patients were allocated to the LPE group and received hCG on the same day. The remaining 31 patients (LPL group) were triggered 24 hours later. All eligible patients were randomised and could be allocated (Figure 1). The data were analysed after correcting for duplicate patients in each group.

No important differences were observed between the LPE and LPL groups and between the HPE and HPL groups for all controlled variables (Table 1): age of the female, gravidity, parity, duration of infertility, ovarian reserve determined by AMH, peak estradiol level, number of transferred embryos, diagnostic criteria (such as female pathology, endometriosis, ovarian dysfunction and PCOS), male pathology, stimulation protocol, number of cancellations, number of failed fertilisations and the experience of the physician performing the embryo transfer.

As soon as three follicles have a diameter ≥ 18 mm, further decisions to pinpoint the moment for administering hCG depend on the progesterone level. If the progesterone level is higher than 1 ng/ml, delaying the administration of hCG by 24 hours has no effect on the number of mature oocytes. If the progesterone level is ≤ 1 ng/ml and 30–50% of the follicles have diameters ≥ 18 mm, delaying oocyte maturation by 24 hours is advised.