Immediate post-denudation intracytoplasmic sperm injection is associated with higher fertilization and early cleavage-stage embryo yield: a prospective sibling-oocyte cohort from Libyan ICSI cycles

Abstract

Intracytoplasmic sperm injection (ICSI) has become an integral practice in assisted reproduction technology (ART), but the optimal time interval between oocyte denudation (DN) and ICSI is still not well defined. Libya remains underrepresented in global fertility estimates, and prospective paired data on DN-to-ICSI timing from Libyan IVF practice are limited. We aimed to investigate whether performing ICSI immediately after DN, compared with an intended four-hour delay, affects pregnancy outcomes and early embryology yield. This prospective cohort study was conducted at the Fertility and Reproductive Medicine Center, Beirut Hospital, Benghazi and included 88 Libyan ICSI cycles. After enzymatic-mechanical DN using hyaluronidase, mature metaphase II (MII) oocytes were injected with selected sperm. Within each cycle, sibling MII oocytes were allocated in a 1:1 ratio to immediate ICSI or delayed ICSI after an intended interval of about four hours, using a computer-generated within-cycle allocation sequence. For clinical comparison, cycles were allocated by pre-generated permuted-block randomization to transfer blastocysts under one timing condition (44 cycles/arm). The primary outcome was normal fertilization (two pronuclei); secondary outcomes included cleavage (Day 3), blastulation and blastocyst quality (Day 5/6; Gardner-Schoolcraft), biochemical pregnancy and clinical pregnancy. Immediate DN–ICSI was significantly associated with higher fertilization rates than delayed injection (80.07% ± 17.09% vs 70.80% ± 17.37%; mean difference: 9.26, 95% CI 4.40–14.12; p = 0.0003) and a higher cleavage rate (76.89% ± 17.03% vs 67.97% ± 18.47%; mean difference 8.93 points, 95% CI 3.59–14.26; p = 0.0013). Biochemical pregnancy (40.9% vs 54.5%; p = 0.200) and clinical pregnancy (29.5% vs 27.3%; p = 0.813) did not differ significantly. These results suggest that a shorter DN-to-ICSI interval may improve fertilization and early embryo development, with no observed differences in blastulation or early pregnancy outcomes. Larger multicenter studies with cumulative live births are required to determine clinical relevance.

1. Introduction

Infertility, a disorder of the male or female reproductive system, is an emerging global health issue. Recent WHO estimates suggest that approximately one in six people of reproductive age experience infertility during their lifetime [1,2]. The Middle East and North Africa (MENA) region, including Libya, has one of the highest global rates of infertility, with total and primary infertility rates of 22.6% and 3.8%, respectively [3]. The WHO global infertility guideline recommends evidence-based prevention, diagnosis, and treatment. It also emphasizes inequities in service quality and the availability of infertility care [4]. In addition, the ICMART world report documented more than 3.5 million ART cycles and over 780,000 infants born. It also emphasizes the importance of optimizing controllable steps within widely used procedures, including ICSI [5].

ICSI is a micromanipulation technique where a single sperm is injected directly into the cytoplasm of a M II oocyte. During ICSI, oocytes undergo micromanipulations such as denudation (removal of the cumulus corona layer) to evaluate maturity and prepare for sperm injection. DN also initiates the DN–ICSI interval, defined as the time from completion of DN to the start of ICSI. During this period, in vitro post-ovulatory aging of M-II oocytes might occur [6]. Therefore, good practice IVF laboratory guidelines emphasize tight control of handling conditions and standardization of key workflow steps, including the DN–ICSI interval [6]. However, available evidence on DN–ICSI interval remains limited and conflicting. Some retrospective cohorts reported associations between timing and clinical pregnancy or live birth [7]. However, other studies including those that used automated time recording, reported no consistent association with fertilization, blastulation, or delivery [8]. Because timing is defined inconsistently across studies, it remains difficult to determine which window is clinically relevant and can be applied in routine practice.

In vitro, post-ovulatory oocyte ageing is a biologically important pathway by which the DN–ICSI interval may affect the outcomes. Following retrieval of M II oocytes, the aging of oocytes is associated with a gradual loss of normal spindle, cytoskeletal instability, mitochondrial dysfunction, dysregulation in calcium homeostasis as well as altered redox potential. These changes may reduce fertilization competence and impair early embryo development [9]. Routine ART manipulations can also increase oxidative stress in vitro, which may interact with procedural timing and affect fertilization and early cleavage dynamics [10]. Therefore, the DN–ICSI interval is a controllable workflow step biologically relevant to routine ICSI practice.

To improve comparability, the IMPRINT initiative adapted CONSORT for infertility trials. In addition, the COMMIT initiative also issued standardized definitions and reporting guidance for core infertility outcomes. However, pregnancy endpoints (biochemical pregnancy, clinical pregnancy, miscarriage, and live birth) are still defined and reported inconsistently across studies, which complicates comparability, evidence, and interpretation [11–13]. These issues may be more pronounced in low- and middle-income countries where financial barriers, limited infrastructure, and fragmented regulation can restrict ART access and limit consistent outcomes [14,15]. In Libya, clinic-based data describe heterogeneous infertility etiologies and support the need for locally generated evidence to guide routine embryology practice and patient counselling. Fertility clinics have only recently been introduced in Libya and their delay was influenced in part by socio-cultural perspectives on ART [16,17]. Prospective evidence on DN-to-ICSI timing remains heterogeneous across timing definitions, study designs, and reported endpoints, and data from Libyan IVF practice remain limited. Therefore, local prospective data are required to inform laboratory workflow and patient counseling.

In this study, we evaluated whether the DN–ICSI interval can influence laboratory and clinical outcomes in routine ICSI practice. We conducted a prospective, single-center sibling-oocyte (split-oocyte) paired study in Benghazi, Libya. Within each cycle, sibling oocytes were prospectively assigned in a 1:1 ratio to immediate post-denudation ICSI or to ICSI after an intended delay of approximately four hours using a computer-generated within-cycle allocation sequence. To reduce cycle-to-cycle variation and improve assessment of laboratory outcomes, we compared sibling oocytes within the same cycle. For clinical outcomes, the embryo-transfer source was assigned at the cycle level to one of the timing arms. The primary outcome measure was normal fertilization, defined as the percentage of injected oocytes with 2PN, and secondary outcome measures included cleavage rate, blastocyst formation rate, embryo morphological quality, pregnancy-related outcomes such as implantation and clinical pregnancy.

2. Materials and methods

2.1. Study design and ethical approval

This was a prospective single-center cohort study conducted between November 2024 and September 2025 at the Fertility and Reproductive Medicine Center, Beirut Hospital, Benghazi, Libya to assess whether the DN–ICSI interval affects embryology and pregnancy outcomes. The laboratory endpoints were evaluated at the within-cycle level in a paired (sibling-oocyte/split-oocyte) design to determine outcomes for ICSI performed immediately post-DN or approximately four hours post-DN. Clinical endpoints were evaluated at the cycle level. Each cycle was allocated to embryo transfer from a single timing arm (44 cycles per arm), using embryos derived from that timing condition. Outcomes assessed were fertilization, cleavage, blastulation and embryo development parameters as well as biochemical and clinical pregnancy. The study cohort and allocation profile are shown in Figure 1. Ethical approval was obtained from the Institutional Review Boards of Beirut Hospital (IRB-BH-FRM/2024/010) in compliance with the Declaration of Helsinki and national data protection regulations. Written informed consent was obtained from all participants prior to the study, and confidentiality was preserved during data collection, analysis and reporting.

Figure 1.

Study cohort and allocation profile of the prospective sibling-oocyte ICSI cohort. A total of 88 ICSI cycles were included in the analysis. Within each cycle, sibling oocytes were prospectively assigned in a 1:1 ratio to immediate and approximately 4-h delayed DN-to-ICSI timing conditions using a computer-generated within-cycle allocation sequence, and embryological outcomes were analyzed as paired within-cycle outcomes. For cycle-level clinical assessment, embryo transfer was evaluated in 44 cycles per arm according to the pre-generated allocation schedule.

2.2. Participants

Couples referred to the Fertility and Reproductive Medicine Center, Beirut Hospital (Benghazi, Libya) for undergoing ICSI were screened consecutively for their eligibility during the study period. We included 88 cycles in which at least one mature oocyte was available for injection, with M-II status verified before ICSI as part of routine laboratory workflow. We excluded cycles with incomplete laboratory records, which may result in the inability to determine days from DN to ICSI and/or calculate other conclusions. Exposure was defined as the time elapsed from oocyte denudation to the start of ICSI. In each included cycle, sibling mature oocytes were prospectively assigned in a 1:1 ratio to one of two predefined denudation-to-ICSI timing conditions: immediate injection after denudation or delayed injection approximately four hours after denudation, using a computer-generated within-cycle allocation sequence. Outcomes were then recorded separately for the two-timing conditions within each cycle. In both conditions, only M II oocytes were injected. A total of 88 ICSI cycles met the inclusion criteria. For clinical comparisons, embryo-transfer source was assigned at the cycle level using a pre-generated permuted-block randomization schedule, resulting in 44 cycles in the immediate-derived transfer arm and 44 cycles in the delayed-derived transfer arm. These denominators (n = 44 per transfer arm; N = 88 total cycles) form the basis of the cycle-level clinical analyses.

2.3. Sample size and power analysis

The sample size was mainly based on logistical consideration of the research plan and estimated clinical volume at the selected recruitment time frame. We therefore recruited 88 ICSI cycles to obtain roughly equal representation of the two DN–ICSI timing exposures for comparative analysis. Since the final sample size was constrained by practical limitations, we examined the effect sizes that this study could detect under a series of standard assumptions (with two-sided α = 0.05 and power equal to 80%) to assess the sensitivity of this design. For binary outcomes in proportions (such as pregnancy outcomes), 44 cycles per group yield ~80% power to detect a moderate-to-large difference between groups (Cohen's h ≈ 0.60). When the base rate is close to midrange, this translates into an absolute difference of approximately 25–30 percentage points. For continuous outcomes comparing two treatment groups at a cycle level (e.g. rates, counts), an assumed sample size of 44 cycles/group provided approximately 80% power to detect a Cohen's d effect size of approximately 0.60. These values are based on an idealized scenario and should not be interpreted as estimates of the true effect size.

2.4. Randomization and blinding

For the within-cycle paired embryology analysis, M II oocytes from the same cycle were prospectively assigned in a 1:1 ratio to one of two DN-to-ICSI timing conditions (immediate post-denudation ICSI or delayed ICSI after an intended interval of four hours) using a computer-generated within-cycle allocation sequence. This approach was used to maintain balance between sibling oocytes allocated to the two timing conditions within each cycle. Laboratory outcomes were analyzed as paired within-cycle comparisons, including fertilization, cleavage, blastulation, and embryo-quality metrics. For the cycle-level clinical comparison, embryo-transfer source was prospectively randomized in a 1:1 ratio to embryos derived from the immediate DN–ICSI condition or the delayed DN–ICSI condition using a pre-generated computer-based permuted-block schedule with varying block sizes prepared before study initiation. Group assignment was applied sequentially at the cycle level in recruitment order, resulting in 44 cycles per transfer arm. This transfer-arm randomization was used only for cycle-level biochemical and clinical pregnancy analyses. Blinding of embryology personnel was not feasible because the intervention involved laboratory timing workflow. Pregnancy outcomes were recorded using the same predefined criteria and follow-up procedures in both transfer arms.

2.5. Clinical protocols

2.5.1. Ovarian stimulation and monitoring

We used the standard protocol of our center using gonadotropins (recombinant follicle-stimulating hormone [(rFSH): Gonal-f® [150–300 IU], Merck Serono, Germany; human menopausal gonadotropin (hMG): Diclair® HP-HMG [150 IU], Germany; and highly purified FSH: Diclar/Diclair® HP-FSH [75 IU], Germany) and GnRH-antagonist (cetrorelix acetate: Cetrotide® [0.25 mg], Merck, Darmstadt, Germany] to induce ovarian stimulation. Baseline assessment was conducted at the start of each cycle and consisted of transvaginal ultrasonography (GE HealthCare; Voluson Expert, Zipf, Austria) to confirm ovarian quiescence as well as baseline pelvic examination. Prestimulation routine included measurement of serum reproductive hormone levels. Gonadotropin baseline levels and ovarian reserve values, serum FSH, luteinizing hormone (LH), and anti-Müllerian hormone (AMH) in serum were also determined using chemiluminescence immunoassay (Mindray, CL 9001, China). Ovarian stimulation was initiated according to individual protocols, with or without addition of hMG. Stimulation began on cycle day 2–3 with rFSH and/or hMG. The initial dose was decided by the treating specialist according to clinical standard criteria such as age, body mass index (BMI), markers of ovarian reserve and previous treatment response when applicable. During stimulation, dose modifications were made according to follicular development as assessed by serial transvaginal ultrasound. FSH and hMG exposure were calculated as cumulative stimulation dosage (doses FSH and HMG). Duration of stimulation measured the time interval from start of treatment to the day of last oocyte maturation trigger. Transvaginal ultrasound scanning (Voluson™ P8, GE HealthCare, Zipf, Austria) was used to follow follicular development and endometrial growth while undergoing ovarian stimulation. Endometrial thickness on the day of trigger was considered as maximum double-layer thickness and simply endo-thickness. Serum estradiol (E2 trigger) and progesterone (P trigger) at the time of triggering was also measured by chemiluminescence immunoassay (Mindray, CL 9001, China). To identify cycle features, the final oocyte maturation was induced by the TH's usual clinical criteria using either urinary hCG (Diclar/Diclair® HP-HCG [5000 IU], Germany) or recombinant hCG (Ovitrelle® [250 mcg], Merck Serono, Italy), as clinically indicated. Oocyte pickup was scheduled after the physician's advice and trigger date.

2.5.2. Oocyte retrieval and DN

Transvaginal ultrasound-guided follicular aspiration with a standard clinical procedure was used to collect oocytes (using a single-use oocyte retrieval needle and aspiration system ((Gynemed GmbH, Germany). Following retrieval, the cumulus–oocyte complexes (COCs) were detected by the embryologist and placed in pre-warmed handling/fertilization medium (Global® Total® w/HEPES, LifeGlobal™, USA) at temperature-controlled conditions until DN was performed. To minimize the fluctuations in temperature and pH, all handling in the laboratory was carried out with warmed surfaces and pre-equilibrated media within an ART/IVF workstation (Esco Medical, Denmark). For oocyte DN, we used a combined enzymatic-mechanical approach. Cumulus cells were dispersed by a short exposure to hyaluronidase (GM501 Hyaluronidase, Gynemed GmbH & Co. KG, Germany), after which gentle mechanical DN was performed with a glass needle (DN/stripping pipette) using the smallest possible opening for residual cumulus and corona removal. Accordingly, we evaluated the nuclear maturation following stripping under an inverted microscope (Narishige Co., Ltd., Tokyo, Japan), and only M II eggs (appearance of first polar body) were selected for ICSI.

We studied the impact of DN-to-ICSI interval on embryological outcomes using two predefined timing criteria. M II oocytes in the immediate group were injected by ICSI immediately following DN in less than an hour post-removal of cumulus-oocyte complexes in hyaluronidase media (GM501 Hyaluronidase, Gynemed, Germany). For the delayed group, M II oocytes were maintained in control culture medium (global® total® w/HEPES, LifeGlobal, USA) and ICSI was performed following a planned four-hour delay. During this time, oocytes were incubated in pre-equilibrated medium according to standard culture conditions. Processing and handling was analogous in both groups. The time of oocyte retrieval, DN and ICSI were recorded in the embryology report. This timing information was applied to derive exposure durations and presented analyses.

2.5.3. Semen preparation and sperm selection

Semen preparation and selection was performed as previously described [17]. Briefly, semen samples were obtained by masturbation in sterile containers on the day of oocyte retrieval and ICSI and were processed using a standard laboratory protocol. Following liquefaction, a semen analysis was performed for routine clinical embryology protocols. Then, the sample was processed for ICSI by sequential wash with sperm handling medium (GM501 SpermAir, Gynemed, Germany). Centrifugation settings were performed as per the hospital standard operating procedure and the relative centrifugation was recorded in the laboratory log to standardize methods within procedures applied. The sperm pellet was resuspended in a new handling medium after the last wash and stored at controlled lab conditions until micromanipulation. For sperm injection (ICSI), the selection of sperm was carried out immediately before injection under an inverted microscope, according to progressive motility and normal morphology. In the ICSI plate, a polyvinylpyrrolidone (PVP) preparation (GM501 PVP 10%, Gynemed, Germany) was added to reduce sperm motility and facilitates controlled capture. Selected spermatozoa were immobilized by a gentle tail nick immediately before aspiration into the injection pipette, following the stereomicroscopic procedure and injected into M II oocytes as detailed below.

2.5.4. ICSI procedure

We carried out ICSI with temperature control on a heated-stage inverted microscope (Narishige Co., Ltd., Tokyo, Japan) at 37 °C using micromanipulators and microinjectors system (Narishige, Japan) within a dedicated ART micromanipulation workstation (Esco Medical, Denmark). Denuded oocytes were manipulated in pre-warmed buffered medium microdroplets under oil (HEPES-buffered handling medium: Global® Total® w/HEPES, LifeGlobal™, USA) to avoid pH and temperature variations during micromanipulation. Only M-II oocytes, which were confirmed under the microscope by a first polar body present at the time of injection, were selected for ICSI. The DN–ICSI interval was performed under the predetermined time conditions mentioned above and all other laboratory manipulation procedures were similar in each condition.

Using a separate micromanipulation workstation, a single immobilized spermatozoan was then aspirated per injection into the injection pipette (Origio/CooperSurgical, Denmark). It was then microinjected into the cytoplasm of the oocyte following the ICSI standard protocol to obtain a uniform, controlled delivery of spermatozoon (zona pellucida-oolemma penetration). Following microinjection, we washed injected oocytes within minutes in fresh pre-equilibrated medium (to remove any residual material) and transferred them to continuous culture in pre-equilibrated embryo culture medium (SAGE 1-Step GM-CSF w/HSA, Origio/CooperSurgical, Denmark). All critical times (date and time of initiation/completion of microinjection) were recorded to evaluate associations between exposure timing and laboratory outcomes.

2.5.5. Embryo culture and assessment

Following ICSI, putative zygotes were washed in fresh, pre-equilibrated culture medium and incubated directly in single-stage embryo culture system (SAGE 1-Step™ GM-CSF with HSA, Origio/CooperSurgical, Denmark). Culture was carried out in pre-equilibrated microdroplets and maintained under controlled incubator conditions (Binder GmbH, Tuttlingen, Germany) according to the hospital routine practice. The embryos were not subjected to ambient conditions except for short, standardized assessment windows and all observations were examined under a heated microscope stage (37 °C) on an inverted microscope (Nikon, Japan) to prevent changes in temperature during the evaluation. Fertilization (2PN and two polar bodies) was assessed at 16–18 h post-injection. Unfertilized zygotes (zygotes that do not meet criteria for normal) were observed, and the number of these was not included in calculation of normally fertilized rate.

The morphological criteria, which included the number of cells, symmetry of blastomeres, and extent of fragmentation were used to identify cleavage stage on day 3. Cleavage-stage embryos matching the top-quality criteria were recorded as a separate subgroup to allow reporting both total cleavage and top-quality cleavages for all stages.

Blastocyst development was assessed on Day 5, and embryos were continued in culture to Day 6 if they had not reached the blastocyst stage on Day 5. Blastocysts were graded according to the Gardner and Schoolcraft classification system (blastocoel expansion stage with inner cell mass and trophectoderm morphology grades). Blastocysts met the predefined top-quality definition were documented separately for reporting of overall blastulation rates as well as top-quality blastulation. The scoring system was not modified, and the time points of evaluation were identical for all timings to ensure that any differences reflected DN–ICSI timing exposure rather than differences in assessment timing.

2.5.6. Embryo transfer and pregnancy evaluation

Embryo transfer was planned at the blastocyst stage and carried out under transabdominal ultrasound guidance (Voluson™ P8, GE HealthCare, Zipf, Austria) in a procedure guided by the routine clinical practice of the clinic. Embryos were selected for transfer according to the timing arm assigned to each cycle allocated to the immediate DN–ICSI timing condition (embryo selection under immediate conditions) or four-hour delay treatment condition (embryo selection under 4-h delayed conditions). The number of embryos transferred was decided in accordance with clinic policy and individual patient counselling. All transfers and immediate peri-transfer details prospectively recorded in the clinical record.

Pregnancy rates were cycle-based, including in clinic follow-up. Biochemical pregnancy was considered as positive β-chorionic gonadotropin (hCG) presence in the serum after embryo transfer. Clinical pregnancy was identified using sonographically confirmed intrauterine gestational sac on ultrasound follow-up. Pregnancy outcome ascertainment was applied similarly across the groups.

2.6. Outcomes and definitions

All definitions of endpoints, outcomes and assessment time points were predefined to ensure that laboratory and clinical endpoints were consistently uniform and are directly related to our results. Laboratory results were characterized at oocyte/embryo level and summarized across cycles to rates or counts with uniform denominators. We included laboratory data such as fertilization, cleavage and blastulation. The cleavage stage and blastocyst developmental parameters were also recorded. In addition, clinical outcomes were considered at the cycle-level where data after embryo transfer was available. Normal fertilization was evaluated on Day 1 after ICSI by observation of zygotes with 2PN and two polar bodies (2PN/2PB) during fertilization check window. Fertilization rate was calculated per cycle by the number of normally fertilized (2PN) zygotes divided by the number of injected M-II oocytes and expressed as a percentage. Abnormal fertilization results, when observed, were annotated in all cases following routine laboratory documents and not considered as normal fertilization.

Cleavage stage development was assessed on Day 3 by standard morphological scoring. The cleavage rate was expressed by cycle as the number of normally fertilized (2PN) zygotes that had divided to cleavage stage on Day 3. Morphological characteristics of cleavage-stage embryos were recorded according to standard morphological criteria and a standardized grading system, which included numbers of cells, symmetry/fragmentation/multinucleation of blastomeres. Blastocyst development was evaluated on Day 5, and observation continued through the Day 6 in cases where blastocysts developed occurred beyond Day 5. The blastulation rate was calculated per cycle as the percentage of 2PN embryos that had developed into a blastocyst by Day 5/6, as reported. Blastocyst morphology was assessed according to the Gardner and Schoolcraft system, with assignment of a predefined top-quality blastocyst category in relation to an embryology unit threshold applied on blastocyst expansion, inner cell mass and trophectoderm quality. These blastocyst characteristics were averaged per cycle and reported as shown in the results section. When embryo transfer was established, biochemical pregnancy was determined as a positive serum hCG following transfer according to the clinic's standard follow-up timetable. Clinical pregnancy was confirmed by ultrasound detecting an intrauterine gestational sac at follow-up. Pregnancy results were consistently recorded according to the same definitions among all groups and summarized per cycle in the results.

2.7. Statistical analysis

Statistical analyses were conducted using the IBM SPSS Statistics (version 27, IBM Corp., Armonk, NY, USA). The complete dataset included 88 ICSI procedures. The data were analyzed according to the nature of the evaluated variables. Laboratory embryology outcome data were obtained for each DN–ICSI timing condition within a cycle (paired within-cycle analyses), whereas clinical pregnancy outcome was analyzed at cycle level in relation to the embryo transfer assignment (both arms provided 44 transfers). Continuous variables are expressed as mean ± SD when approximately normally distributed, and median (interquartile range) if skewed; categorical variables are shown as number (percentage). Normality assumptions were assessed through Shapiro–Wilk and one-sample Kolmogorov–Smirnov tests, as well as histograms and QQ plots. All tests of hypotheses were two-sided and p-values less than 0.05 were considered to be statistically significant. To compare immediate ICSI with four-hour delay within a cycle, paired-samples t-tests were applied where the differences between paired observations were normally distributed. When this assumption was violated, the Wilcoxon signed rank test was used. Paired effects are presented as the overall mean difference within pairs (95% confidence interval), and standardized paired effect sizes when appropriate. Laboratory rates and associated denominators were calculated based on standard IVF key performance indicator (KPI) definitions. For instance, fertilization rate was calculated as the proportion of 2PN zygotes to injected M II oocytes.

The two transfer groups were compared on baseline characteristics and clinical outcomes. For continuous variables, we used an independent samples t-test when appropriate for assumptions. If the conditions were not met, we used the Mann–Whitney U test. Pearson's chi-square test was employed for categorical measures including biochemical and clinical pregnancy. The Fisher's exact test was used if expected cell numbers were small. Clinically relevant effects are presented with 95% confidence intervals and with absolute risk differences, when appropriate relative measures of the effect are also reported. The statistical associations among continuous variables were examined using the Pearson correlation coefficients with 95% confidence intervals and p-value. Independent predictors of biochemical and clinical pregnancy for the cycle level were identified using multivariable binary logistic regression.

In addition, we conducted sensitivity analyses to evaluate the robustness of our findings. For the primary paired embryology comparisons, the nonparametric Wilcoxon signed-rank test was also used. Correlation analyses were examined using Spearman's rank order coefficients to reduce sensitivity to outliers and non-linear associations. For logistic regression, uncertainty was assessed using bootstrap confidence intervals and effect diagnostics were tested to confirm that the findings were not caused by a small set of cycles. Correction for multiple testing was not conducted because outcomes were pre-specified, and results were reported in terms of effect sizes with confidence intervals.

3. Results

3.1. Participant characteristics

A total of 88 ICSI cycles were included in the study. Laboratory embryology outcomes were evaluated as paired within-cycle comparisons after prospective assignment of sibling oocytes to immediate or delayed DN-to-ICSI timing conditions. For cycle-level clinical comparisons, embryo transfer was evaluated in 44 cycles per arm. Their baseline characteristics are presented in Table 1. The mean female age was 28.41 ± 4.44 years, while men were 33.08 ± 4.45 years old. Mean height was 1.59 ± 0.06 m, weight was 71.69 ± 9.05 kg, BMI was 28.46 ± 4.47 kg/m². The mean duration of infertility was 4.64 ± 1.67 (in the range of 2–8 years). We tested for normality (bell-shape) in the numeric data using two tests (Kolmogorov–Smirnov and Shapiro–Wilk). Female age, male age, height, weight, and infertility duration did not follow a normal distribution. BMI was approximately normally distributed. The two groups were comparable at baseline. The mean age of females in immediate DN group and four-hour delay group were 28.64 ± 4.70 years and 28.18 ± 4.22 years respectively. BMI was 28.69 ± 4.36 kg/m² and 28.22 ± 4.62 kg/m², respectively. The proportion of patients with primary and secondary infertility was 67.0% (59/88) and 33.0% (29/88), respectively. Correspondingly, 73.9% (65/88) were childless, 25.0% (22/88) had one child and 1.1% (1/88) had three children.

Table 1.

Participant demographics and baseline reproductive variables (N = 88).

Variable	Mean ± SD	Median [IQR]	Range (min–max)	95% CI for mean	Normality p (K-S/S-W)
Age (wife), years	28.41 ± 4.44	28.00 [7.00]	20.00–40.00	27.47–29.35	0.025
Age (husband), years	33.08 ± 4.45	32.50 [6.00]	23.00–43.00	32.14–34.02	0.041
Height, m	1.59 ± 0.06	1.59 [0.09]	1.49–1.72	1.58–1.60	0.010
Weight, kg	71.69 ± 9.05	71.50 [15.75]	57.00–86.00	69.78–73.61	<0.001
BMI, kg/m²	28.46 ± 4.47	27.91 [7.03]	20.44–38.29	27.51–29.40	0.061
Infertility duration, years	4.64 ± 1.67	5.00 [3.00]	2.00–8.00	4.28–4.99	<0.001
Infertility type	Primary: 59 (67.0%); secondary: 29 (33.0%)	–	–	–	–
Parity (children)	0: 65 (73.9%); 1: 22 (25.0%); 3: 1 (1.1%)	–	–	–	–

Values are presented as mean ± SD, median [IQR], range (min–max), and 95% CI for the mean, where applicable. Normality was assessed using the Kolmogorov–Smirnov and Shapiro–Wilk tests. For each variable, the smaller p-value from the two tests is reported.

3.2. Baseline semen parameters

Seminal characteristics at baseline were analyzed in all 88 subjects. The mean of semen volume was 2.80 ± 0.92 mL (95% CI: 2.60–2.99). The mean sperm concentration was 38.92 ± 15.99 × 10⁶/ml (95% CI: 35.58–42.26). The percentage of mean progressive motility was 14.32% ± 6.96% (95% CI: 12.86%–15.77%) and the percentage of mean non-progressive motility was 19.55% ± 6.19% (95% CI: 18.25%–20.84%). The total motile count (TMC) and the mean of non-motile sperm index (SA-NF) was 13.39 ± 7.91 × 10⁶ (95% CI: 11.74–15.04 × 10⁶) and 0.65 ± 0.71, respectively (95% CI: 0.50–0.80). Normal distribution was verified by one-sample Kolmogorov–Smirnov test. Using the one-sample Kolmogorov–Smirnov test, semen volume (D = 0.247, p < 0.001), progressive motility (D = 0.221, p < 0.001), non-progressive motility (D = 0.232, p < 0.001), total motile count (D = 0.157, p < 0.001), and SA-NF (D = 0.307, p < 0.001) did not follow a normal distribution. In contrast, sperm concentration did not show a significant deviation from normality (D = 0.085, p = 0.158).The baseline characteristics of the semen parameters are shown in Table S1.

3.3. Ovarian stimulation and cycle characteristics

Baseline ovarian stimulation and cycle characteristics were measured for all 88 patients. Normality of the data was examined by one-sample Kolmogorov–Smirnov test with Lilliefors correction (Table S2). The distribution of the baseline hormone values (FSH, LH and AMH) was not found to significantly deviate from normality for any of them ((FSH: D = 0.066, p = 0.200; LH: D = 0.083, p = 0.192; AMH: D = 0.066, p = 0.200). Estradiol on trigger day (E2-trigger) also showed no significant deviation from normality (D = 0.093, p = 0.060). In contrast, variables associated with gonadotropin dosing and ovarian response deviated from normality including total FSH (FSH: D = 0.157, p < 0.001), total hMG dose (HMG: D = 0.232, p < 0.001) and stimulation duration (stimulation days: D = 0.198, p < 0.001). On trigger day, P-trigger was also not normally distributed (D = 0.096, p = 0.044). Endometrial thickness at the day of trigger (Endothickness: D = 0.312, p < 0.001), number of cumulus-oocyte complexes retrieved (COCs: D = 0.144, p < 0.001) and counts of M II oocytes (M II-A: D = 0.178, p < 0.001; M II-B: D = 0.213, p < 0.001) were similarly non-normally distributed. We used these distributional findings to define the appropriate statistical framework for subsequent outcome comparisons.

3.4. Fertilization outcomes: comparison between immediate DN and four-hour delay

We evaluated fertilization in all 88 ICSI cycles by comparing the two DN procedures in a single cycle (paired analysis). The average fertilization rate was 80.07% ± 17.09% for the immediate DN group and 70.80% ± 17.37% for the delayed four-hour group (Table 2). The paired-samples test showed a significant decrease in fertilization with the four-hour delay (mean difference = 9.26; 95% CI: 4.40–14.12; t = 3.787; p = 0.0003). The pooled effect size was a Cohen's dz = 0.40 (95% CI: 0.19–0.62). Estimates for analyses presented are unadjusted comparisons.

Table 2.

Fertilization rate by denudation condition (paired within-cycle data) and paired comparison (n = 88 ICSI cycles).

Outcome	Immediate denudation (mean ± SD, %)	Four-hour delay (mean ± SD, %)	Mean difference (%)	95% CI (difference)	t (df)	p-value	Effect size (Cohen's dz)	95% CI (dz)
Fertilization rate	80.07 ± 17.09	70.80 ± 17.37	9.26	4.40 to 14.12	3.787 (87)	0.0003*	0.40	0.19 to 0.62

Values are mean ± SD unless stated. Mean difference was calculated by subtracting the 4-h delay value from the immediate-denudation value. Abbreviations: CI, confidence interval; df, degrees of freedom; SD, standard deviation; dz, standardized paired effect size (Cohen's dz). Significant p-values are bolded and marked with an asterisk (*).

3.5. Cleavage (division) outcomes

Cleavage (division) results were evaluated in all 88 ICSI cycles on a paired, within-cycle basis comparing oocytes denuded immediately and those denuded after a four hours interval. The average cleavage rate was 76.89% ± 17.03% for immediate DN (n = 88) and 67.97% ± 18.47% when there was a delay of four hours as presented in Table 3. The paired-samples t-test showed a significantly higher cleavage rate with immediate DN, with a mean paired difference of 8.93 percentage points (95% CI: 3.59 to 14.26), t(87) = 3.327, p = 0.0013. Cohen's dz of the standardized paired effect size was 0.35 (95% CI: 0.14 to 0.57).

Table 3.

Cleavage (division) rate by denudation condition and paired within-cycle comparison (N = 88 paired cycles).

Outcome	Immediate denudation (mean ± SD, %)	95% CI for mean (%)	Four-hour delay (mean ± SD, %)	95% CI for mean (%)	Mean paired difference (%)	95% CI of difference (%)	t (df)	p-value	Standardized paired effect size (Cohen's dz, 95% CI)
Cleavage (division) rate	76.89 ± 17.03	73.28 to 80.50	67.97 ± 18.47	64.05 to 71.88	8.93	3.59 to 14.26	3.327 (87)	0.0013*	0.35 (0.14 to 0.57)

Values are mean ± SD unless stated. The paired mean difference was calculated by subtracting the 4-h delay value from the immediate-denudation value. Abbreviations: CI, confidence interval; df, degrees of freedom; SD, standard deviation. Significant p-values are bolded and marked with an asterisk (*).

3.6. Blastulation outcomes: immediate DN versus four-hour delay

The two DN conditions were compared within the cycle of processing using a paired, within-cycle analysis of blastulation outcomes from the 88 ICSI cycles. Mean blastulation rates were 63.01% ± 20.49% following immediate DN and 62.50% ± 13.27% for the four-hour delay (Table 4). The paired-samples t-test did not show a significant difference by these condition (mean difference = 0.51 percentage points; 95% CI: −4.93 to 5.94; t = 0.186; p = .853. The standardized paired effect size was negligible (Cohen's dz = 0.02; 95% CI, −0.19 to 0.23).

Table 4.

Blastulation (blastocyst formation) rate by denudation condition and paired within-cycle comparison (N = 88 paired cycles).

Outcome	Immediate denudation (mean ± SD, %)	95% CI (mean)	Four-hour delay (mean ± SD, %)	95% CI (mean)	Mean difference (%)	95% CI (difference)	t (df)	p-value	Effect size (dz)	95% CI (dz)
Blastulation rate	63.01 ± 20.49	58.67–67.35	62.50 ± 13.27	59.69–65.31	0.51	−4.93 to 5.94	0.186 (87)	0.853	0.02	−0.19 to 0.23

Values are mean ± SD with 95% CI for condition means. Mean difference was calculated by subtracting the 4-h delay value from the immediate-denudation value. Abbreviations: CI, confidence interval; df, degrees of freedom; SD, standard deviation; dz, Cohen's dz. Significant p-values are bolded and marked with an asterisk (*).

3.7. Embryo quality outcomes: day 3 and day 5 comparisons

We evaluated the embryo development on day 3 and day 5/6 in the same cycle (n = 88 ICSI cycles) to compare the two DN conditions as shown in Table 5. We found that on Day 3, the mean number of cleavage-stage embryos was 4.23 ± 1.45 with immediate DN compared to 3.74 ± 2.40 with a four-hour delay. The paired t-test showed a statistically significant difference (mean diff = 0.49; 95% CI: 0.04 to 0.94; p = 0.0349), with a small, paired effect size (Cohen's dz = 0.23; 95% CI: 0.02 to.44). In contrast, the mean number of top blastocysts on Day 5 (D5/6 Top) was 1.74 ± 1.06 with immediate DN and 1.32 ± 0.89 with a four-hour delay. This difference was also statistically significant (mean difference = 0.42; 95% CI: 0.16 to 0.68; t(87) = 3.234; p = 0.0017), with a small-to-moderate paired effect (Cohen's dz = 0.34; 95% CI: 0.13 to 0.56).

Table 5.

Day 3 and Day 5/6 embryo metrics by timing condition with within-cycle paired comparisons (N = 88 paired cycles).

Outcome	Condition	Mean ± SD	95% CI (mean)	Paired comparison (test)	Statistic (s)	p-value	Effect size	95% CI (effect size)
Day 3 cleaved embryos (D2/3_No.)	Immediate denudation	4.23 ± 1.45	3.92–4.53	Paired t-test	t(87) = 2.143; mean diff = 0.49 (0.04 to 0.94)	0.0349*	Cohen's dz = 0.23	0.02 to 0.44
	four-hour delay	3.74 ± 2.40	3.23–4.25
Day 5/6 top blastocysts (D5/6 Top)	Immediate denudation	1.74 ± 1.06	1.51–1.96	Paired t-test	t(87) = 3.234; mean diff = 0.42 (0.16 to 0.68)	0.0017*	Cohen's dz = 0.34	0.13 to 0.56
	four-hour delay	1.32 ± 0.89	1.13–1.51

Values are mean ± SD with 95% confidence intervals (CI) for the mean. Paired comparisons are based on within-cycle differences (paired t-test). Effect size is reported as Cohen's dz with 95% CI. Abbreviations: D2/3-No., number of cleavage-stage embryos assessed on Day 2/3 (Day 3); D5/6 Top, number of top-quality blastocysts assessed on Day 5/6; SD, standard deviation; CI, confidence interval. Significant p-values are bolded and marked with an asterisk (*). p < 0.05.

3.8. Pregnancy outcomes: biochemical and clinical confirmation

Pregnancy rates were evaluated per cycle in 88 ICSI cycles: 44 in the immediate DN group and 44 in the delay for four hours group. Biochemical pregnancy (β-hCG positivity) was observed in 18/44 cycles (40.9%) in immediate DN and 24/44 cycles (54.5%) in four-hour delay groups. There was no significant difference between the groups (Pearson χ²(1) = 1.640, p = 0.200; φ = 0.14). Clinical pregnancy, defined as the presence of an intrauterine gestational sac on ultrasound, was observed in 13/44 cycles (29.5%) in the immediate DN group and 12/44 cycles (27.3%) in the four-hour delay group. Again, no significant difference was detected between the groups (Pearson χ²(1) = 0.056, p = 0.813; Fisher's exact two-sided p = 1.000; φ = 0.03). The pregnancy outcome data, along with Wilson 95% confidence intervals for each proportion, are summarized in Table 6.

Table 6.

Pregnancy outcomes by group with χ²/Fisher tests (N = 88).

Outcome	Group	Positive n (%)	Negative n (%)	95% CI (Wilson) for % positive	Test	χ² (df)	p-value	Fisher p (two-sided)	φ
Biochemical pregnancy	Immediate denudation (n = 44)	18 (40.9%)	26 (59.1%)	27.7%–55.6%	Pearson χ²	1.640 (1)	0.200	–	0.14
	four-hour delay (n = 44)	24 (54.5%)	20 (45.5%)	40.1%–68.3%
Clinical pregnancy	Immediate denudation (n = 44)	13 (29.5%)	31 (70.5%)	18.2%–44.2%	Pearson χ²	0.056 (1)	0.813	1.000	0.03
	four-hour delay (n = 44)	12 (27.3%)	32 (72.7%)	16.3%–41.8%

Values are counts (percentages). Exact test statistics and p-values are shown. Abbreviations: CI, confidence interval; df, degrees of freedom; φ, phi coefficient. Wilson score intervals are shown for the proportion positive in each cell.

3.9. Correlation analyses

Pearson correlations were carried out to examine linear relationships between different baseline parameters and semen parameters in the entire cohort (n = 88). The correlation results are summarized in Figure 2. The female age and male age were positively correlated, r(86) = 0.90, 95% CI [0.85, 0.93], p < 0.001. There was also a strong positive correlation between weight and BMI, r(86) = 0.89, 95% CI [0.84, 0.93], p < 0.001. For semen parameters, sperm concentration was strongly positively correlated with total motile count, r(86) = 0.80, 95% CI [0.71, 0.87], p < 0.001. Non-progressive motility was also positively associated with total motile count, r(86) = 0.60, 95% CI [0.45, 0.72], p < 0.001. In contrast, the SA-NF had a negative relation to total motile count, r(86) = −0.37, 9.

Figure 2.

Forest plot of Pearson's correlation coefficients (N = 88). Points denote Pearson's r and whiskers the 95% confidence intervals (Fisher z). The vertical reference line denotes no correlation (r = 0); values to the right indicate positive correlations and to the left negative correlations. Two-sided p-values are shown on the right of each row. Abbreviations: BMI, body mass index; CI, confidence interval; SA-NF, non-motile sperm metric; TMC, total motile count.

3.10. Logistic regression models

To test whether age, duration of infertility, AMH, and BMI were associated with biochemical and clinical pregnancies in the cohort (N = 88), we used multivariable binary logistic regression models. Adjusted odds ratios (AORs) with 95% confidence intervals are reported in Table S3 and summarized in Figure 3. For biochemical pregnancy, we found that all predictors were statistically non-significant including age (AOR 1.06, 95% CI 0.96–1.18, p = 0.258), infertility duration (AOR 0.99, 95% CI 0.76–1.29, p = 0.938), AMH (AOR 0.88, 95% CI 0.49–1.60, p = 0.683), and BMI (AOR 1.00, 95% CI 0.91–1.10, p = 0.949). Overall, the model did not provide a statistically significant improvement over an intercept-only model (χ²(4) = 1.993, p = 0.737). The amount of variation explained was very small, with pseudo-R² values of 0.022 (Cox and Snell) and 0.030 (Nagelkerke). For clinical pregnancy, none of the variables in the model showed a meaningful or statistically significant association with the outcome: age (AOR 1.08, 95% CI 0.96–1.21, p = 0.180), duration of infertility (AOR 0.99, 95% CI 0.74–1.33, p = 0.942), AMH (AOR 1.09, 95% CI 0.57–2.08, p = 0.798), and BMI (AOR 0.98, 95% CI 0.88–1.09, p = 0.734). Taken together, the model was not statistically significant (χ²(4) = 2.233, p = 0.693) and accounted for only a small variation in clinical pregnancy (pseudo R-squared = 0.025 by Cox and Snell; 0.036 Fig.by Nagelkerke). Using a 0.50 probability cut-off, the model showed an overall accuracy of 71.6%, but it did not identify any clinical pregnancies (0.0% sensitivity) and correctly classified all non-pregnancy cycles (100.0% specificity).

Figure 3.

Adjusted odds ratios (OR) with 95% CIs from logistic regression models. Forest plot displays ORs (points) with 95% CIs (horizontal bars) for biochemical and clinical pregnancy models; right-side labels show exact p-values. Abbreviations: AMH, anti-Müllerian hormone; BMI, body mass index; CI, confidence interval; OR odds ratio.

4. Discussion

The DN-to-ICSI interval is a clinically relevant parameter, which may affect oocyte competence during this brief period post-retrieval. During this time window, temporally dependent cell changes can accumulate. The oocyte might become more vulnerable to perturbations in its environment following cumulus removal, e.g. temperature, pH and subsequent manipulations. If the oocyte remains in vitro for a long period of time, postovulatory aging-related consequences can accelerate [9,10]. This can decrease the fertilization ability, and even slow the early cleavage kinetics. Therefore, Good laboratory practice requires a steady workflow, avoidance of unnecessary delays, and stable micromanipulation conditions [6,18].

In this cohort, immediate post-denudation ICSI was associated with higher fertilization rates and higher cleavage rates, together with modest increases in the number of cleavage-stage embryos and top-quality blastocysts per cycle. In contrast, blastulation rate and cycle-level biochemical and clinical pregnancy outcomes did not differ significantly between the two timing conditions. These findings suggest that shortening the DN-to-ICSI interval may improve early laboratory outcomes, whereas any effect on early pregnancy outcomes was not detectable in this cohort.

The observed patterns, marked differences at fertilization and cleavage with convergence out by blastocyst stage and in pregnancy outcomes, was consistent with the published evidence suggesting that timing of DN and injection relative to that ovulation may impact the early embryo competence [19–22]. However, there are some other factors that can influence pregnancy outcomes, including how embryos are selected, uterine and endometrial factors, as well as limited statistical power for pregnancy-related endpoints in many cohorts. Cohort based data also suggested that reducing the duration of time between meiotic DN and ICSI may be of advantage. Recent large retrospective studies also demonstrate that the interval between DN and injection may be associated with clinical outcomes, but direct cross-study comparison is constrained by heterogeneity as the definition of this interval, stimulation protocol and embryo scoring systems were not standardized [19–22].

Immediate DN is biologically plausible because a short post-DN interval may reduce time-dependent oocyte deterioration such as instability of the meiotic spindle, increased oxidative stress and impairment of mitochondrial function. Consequently, these alterations may reduce the efficiency of fertilization and interfere with the timing and synchronism of early cleavages. This is consistent with the observation that blastocyst quality and pregnancy outcomes were similar between groups. The early developmental loss that occurs may be considered a biological filter. These laboratory observations remain relevant, even if they did not translate into a detectable clinical difference while those with sufficient developmental competence advance to transfer. These results are consistent with the perspective that laboratory findings should be assessed in the context of whole clinical course or standardized laboratory conditions [19,23,24].

From a clinical standpoint, these findings suggest that better early laboratory outcomes, including fertilization and cleavage, were not associated with an obvious difference in pregnancy rates in this group. However, these laboratory observations remain relevant, even if it did not translate into a detectable clinical difference. These findings suggest that other factors that influence pregnancy were not measured or cannot be changed and embryo selection at later stages may narrow early differences by the time of transfer. Our regression results support this interpretation. The regression models did not identify significant associations of age, infertility duration, AMH, or BMI with biochemical or clinical pregnancy, and overall model distinction was limited. These findings suggested that the timing threshold alone may have little clinical application. Thus, ART add-ons emphasized that the interpretation should be based on patient-important outcomes such as live birth and safety in the context of evidence resulted from laboratory measures without improvement in clinical outcomes [25].

Our findings suggest that the DN–ICSI interval is a modifiable laboratory step. The aim is to avoid unnecessary time delays and maintain stable conditions during DN/injection. Any changes should remain within established good-practice and governance standards, regular monitoring of procedures, and periodic review of key performance parameters [18,26]. Furthermore, we reported stage-specific aspects to investigate differences across early kinetics, later embryo development, and pregnancy outcomes. In addition, we reported effect sizes (with confidence intervals) alongside p-values to show the size and precision of effects, not only whether they reach statistical significance. In this cohort, the multivariable models did not identify significant predictors of biochemical or clinical pregnancy, and their discriminatory performance was limited.

Our study has some limitations. We conducted the experiments in a single center and thus results may not be applied to all IVF laboratories because laboratory practices and definitions are not consistent between different IVF centers. In our study, it was not feasible to blind the exposure because the timing step was itself the exposure. Thus, slight variations in manipulations or assessments cannot be ruled out when the same scoring method and predetermined timing are used. The sibling-oocyte approach provides an internally controlled comparison of laboratory results because oocytes are compared within the same cycle. Although oocytes retrieved from the same patient and are handled by the same team, patient and cycle level influences may persist. This sample size was more suitable for detecting paired within-cycle differences in embryology outcomes than for detecting small differences in biochemical or clinical pregnancy. Accordingly, the absence of a significant difference in pregnancy outcomes should be interpreted cautiously and should not be taken as evidence of equivalence between the two-timing strategies. Finally, follow-up was limited to biochemical and clinical pregnancy rather than live birth, and hence the findings are not directly applicable to live birth rates. Future studies should employ multicenter, larger, prospective investigations using consistent definitions of timing intervals. Confounding factors should be considered and reported with good baseline comparability among the studies, while outcomes should be reported in accordance to standardized core outcome frameworks [27,28]. Adopting infertility core outcome sets and consistent outcomes definitions will improve comparability between studies. Evidence will be improved if data are stratified according to relevant clinical and laboratory contexts (e.g. ovarian response, stage of embryo transfer, culture system), to determine where timing effects might be most clinically relevant [27,28].

5. Conclusion

The present prospective sibling-oocyte ICSI study demonstrates that performing ICSI immediately after oocyte DN resulted in improved fertilization and cleavage rates compared with four hours post-DN delay. In contrast, blastocyst formation was not different between both groups. We did not observe significant differences in biochemical and clinical pregnancy rates between the two approaches. This finding suggests that possible effects of DN–ICSI timing on clinical outcomes are probably affected by other biological or patient factors or not readily detectable from this sample size. Overall, these findings are consistent with IVF laboratory good-practice guidance supporting minimization of unnecessary delays between DN and ICSI may result in improved early embryological outcomes. Our findings may require further validation by larger clinical and laboratory multicenter studies to define how the DN–ICSI timing should be controlled in clinical IVF practice to improve fertility care in the future.

Supplementary Material

Supplementary Material

LJM_Supplementary Information_V2.docx

Disclosure statement

No potential conflict of interest was reported by the author(s).

Funding

This research did not receive any external funding or institutional support. All expenses related to the study, including laboratory procedures, data analysis, and manuscript preparation, were fully financed by the authors.

Data availability statement

The data supporting this study's findings are available from the corresponding author upon reasonable request. Due to ethical and privacy constraints related to human subjects, raw datasets are not publicly accessible. Data sharing will comply with institutional data protection policies and applicable regulations.

Supplemental material

Supplemental data for this article can be accessed at https://doi.org/10.1080/19932820.2026.2666920.