Abstract
Purpose
We aimed to compare embryo development, cumulative live birth rate (CLBR), and perinatal outcomes of embryos cultured in 20% and 5% oxygen from days 1 to 3 after insemination.Methods
This retrospective study included patients who received in vitro fertilization (IVF) treatment between January 2015 and November 2019. Embryos of each patient were cultured at 20% or 5% oxygen from days 1-3 after insemination. The primary outcome was CLBR. Propensity score matching (PSM) was used to balance patients' baseline data in both oxygen groups.Results
In total, 31,566 patients were enrolled. After PSM, the rate of high-quality day 3 embryos was significantly lower in the 20% than in the 5% oxygen group (0.49 ± 0.33 vs 0.51 ± 0.33; adjusted β = -0.03; 95% confidence interval [CI], -0.03 to -0.02). The CLBR was significantly lower in the 20% than in the 5% oxygen group (58.6% vs. 62.4%; adjusted odds ratio = 0.85; 95% CI, 0.81-0.90). The birthweight and Z score of singletons were significantly higher in the 20% than in the 5% oxygen group (birthweight: 3.30 ± 0.50 vs. 3.28 ± 0.48; adjusted β = 0.022; 95% CI, 0.004-0.040; Z score: 0.26 ± 1.04 vs. 0.22 ± 1.01; adjusted β = 0.037; 95% CI, 0.001-0.074).Conclusion
Culturing embryos at atmospheric oxygen concentrations from days 1 to 3 compromises embryo quality, reduces CLBR, and affects birthweight. The 5% oxygen concentration is more suitable for embryo culture in IVF laboratories to achieve successful outcomes.Free full text
Oxygen concentration from days 1 to 3 after insemination affects the embryo culture quality, cumulative live birth rate, and perinatal outcomes
Abstract
Purpose
We aimed to compare embryo development, cumulative live birth rate (CLBR), and perinatal outcomes of embryos cultured in 20% and 5% oxygen from days 1 to 3 after insemination.
Methods
This retrospective study included patients who received in vitro fertilization (IVF) treatment between January 2015 and November 2019. Embryos of each patient were cultured at 20% or 5% oxygen from days 1–3 after insemination. The primary outcome was CLBR. Propensity score matching (PSM) was used to balance patients’ baseline data in both oxygen groups.
Results
In total, 31,566 patients were enrolled. After PSM, the rate of high-quality day 3 embryos was significantly lower in the 20% than in the 5% oxygen group (0.49 ± 0.33 vs 0.51 ± 0.33; adjusted β = −0.03; 95% confidence interval [CI], −0.03 to −0.02). The CLBR was significantly lower in the 20% than in the 5% oxygen group (58.6% vs. 62.4%; adjusted odds ratio = 0.85; 95% CI, 0.81–0.90). The birthweight and Z score of singletons were significantly higher in the 20% than in the 5% oxygen group (birthweight: 3.30 ± 0.50 vs. 3.28 ± 0.48; adjusted β = 0.022; 95% CI, 0.004–0.040; Z score: 0.26 ± 1.04 vs. 0.22 ± 1.01; adjusted β = 0.037; 95% CI, 0.001–0.074).
Conclusion
Culturing embryos at atmospheric oxygen concentrations from days 1 to 3 compromises embryo quality, reduces CLBR, and affects birthweight. The 5% oxygen concentration is more suitable for embryo culture in IVF laboratories to achieve successful outcomes.
Supplementary Information
The online version contains supplementary material available at 10.1007/s10815-023-02943-4.
Introduction
Human embryos are exposed to physical or chemical factors such as pH, temperature, oxygen concentration, and volatile organic compounds during in vitro culture. These factors, particularly oxygen concentration, are critical to regulating embryo development [1]. The physiological oxygen level in the oviduct and uterus of mammals is between 2 and 8% [2]. In 1971, Patrick Steptoe and Robert Edwards successfully cultured human embryos to the blastocyst stage in a low-oxygen environment (5%) [3]. In 1981, they reported that the transfer of embryos cultured at atmospheric oxygen concentrations (20%) could also lead to successful pregnancies [4]. Embryo culture under hypoxia requires a unique three-gas culture system (adding nitrogen) and oxygen sensor. Therefore, dual-gas incubators comprising a mixture of air and carbon dioxide were widely used during the first 20 years of in vitro fertilization (IVF) to avoid additional costs. A worldwide survey conducted in 2014 showed that among the respondents, 39% used only 20% oxygen culture, 24% used only 5% oxygen culture, and 34% used both 20% and 5% oxygen culture [5].
During animal research, excessive oxygen produces more reactive oxygen species (ROS) by high-energy electrons “leaking” along the electron transport chain during oxidative phosphorylation, which has toxic effects on embryo development [6, 7]. Some clinical studies tracked the effect of oxygen concentration on embryo quality and found that culture with a low oxygen concentration was effective in improving embryo quality [8, 9]. These findings raised concerns about the use of an atmospheric oxygen concentration for embryo culture. Embryo fragmentation and slow development are two main indicators when evaluating embryo quality and predicting pregnancy outcomes. Exploring the relationship between oxygen concentration and these two conditions can provide more detailed reference information for improving laboratory culture conditions.
Some studies have shown that hypoxia could improve the implantation rate and live birth rate [10], whereas others did not find significant differences in clinical measures between groups with 20% and 5% oxygen concentrations [11]. Meta-analyses of the data provided low-quality evidence to substantiate that 5% oxygen concentrations could slightly improve pregnancy outcomes; therefore, the authors were unsure about the difference in this comparison [12]. We speculated that the critical reason for the uncertainty of these results is the study design variations, including endpoint indicators, embryo transfer strategies, inclusion of populations, and sample size. Another reason is that a single transfer does not fully reflect the overall quality of embryos during a cycle. The cumulative live birth rate (CLBR) refers to live births after continuous transfer, including fresh and frozen embryos, which better summarizes the chances of a live birth throughout the treatment cycle [13]. Therefore, CLBR can better reflect the influence of variables in the culture environment, such as the oxygen concentration, on the overall embryo quality. To date, only one study has used CLBR as an endpoint indicator and reported differences in the CLBR of embryos cultured in different oxygen concentrations [14].
During animal research, the atmospheric oxygen concentration affects gene expression [15, 16], DNA methylation [17], protein mass spectrometry [18], and embryo metabolism [19] in mammalian embryos, thus raising concerns about the maternal and fetal risks using atmospheric oxygen concentrations for IVF. Regarding human studies, only two studies tracked perinatal outcomes. Maria et al. followed up the live birth outcomes and neonatal outcomes of a previous randomized controlled trial and determined that the oxygen concentration during a 3-day embryo culture did not affect the neonatal birth weight [20]. Aafke et al. conducted a retrospective analysis and showed similar perinatal outcomes [14]. However, these two studies included no more than a few hundred live births; therefore, more data are needed to draw cogent conclusions.
We aimed to compare the embryo quality and CLBR cultured at 20% and 5% oxygen concentrations based on a large sample size. We also conducted 1 to 5 years of follow-up to compare perinatal outcomes between atmospheric and low oxygen levels.
Methods
Study design and patients
This was a retrospective cohort study. We extracted data from patients in the CITIC-Xiangya Assisted Reproductive Technology Cohort (NCT05404464). We included the first oocyte retrieval cycles from all patients registered between January 1, 2015, and November 30, 2019, with transfer dates limited to November 30, 2020. Eligible couples were diagnosed with subfertility, underwent conventional IVF insemination, and obtained at least one normal fertilized oocyte. Donor sperm, donor oocyte, couples with any abnormal karyotype or monogenic disease, a known uterine malformation (unicornuate uterus, bowed uterus, double uterus, mediastinum, or residual horn uterus), untreated hydrosalpinx, moderate-to-severe uterine adhesion, recurrent miscarriage, and severe oligoasthenospermia were excluded. This study was approved by the Institutional Review Board of Reproductive and Genetic Hospital of CITIC-Xiangya (LL-SC-2020-019).
Stimulation protocol and embryo culture
The ovarian stimulation protocols were performed as described by Tan et al. [21]. When two-thirds of the follicles reached 18 mm on ultrasonography, the patient was injected with hCG (5000–10,000 IU Pregnyl, Merck, Rahway, NJ, USA) to stimulate the ovaries. Cumulus oocyte complexes (COCs) were collected 34 to 36 h after hCG administration using transvaginal ultrasound. All COCs were placed in fertilization medium (G-IVF, Vitrolife, Goteborg, Sweden) at 37°C in an atmosphere of 6% CO2 for 3 to 4 h. Subsequently, 2–3 COCs were transferred to 1 mL G-IVF medium in a fertilization tube (BD Falcon 352003) and co-cultured with motile sperm at a concentration of 1.0 × 105/mL overnight in a 37°C incubator with 6% CO2 and 20% O2. All the fertilization tubes of a given patient were placed in a 50-ml glass beaker (Sichuan Shubo, China). Fertilization was assessed at 16 to 18 h after insemination. Zygotes with two pronuclei (2PN) were identified as normally fertilized and transferred to a cleavage medium (G1.5, Vitrolife). The staff randomly selected a glass beaker from the incubator for fertilization assessment. The zygotes of the first four patients were placed into the atmospheric oxygen concentration (6% CO2 and 20% O2; Thermo Forma 3121, Forma Scientific Inc., Marietta, OH, USA), and the zygotes of the next four patients were placed into the low-oxygen level incubator (6% CO2, 5% O2, and 89% N2; ASTEC APM-50D, ASTEC Co., Ltd., Fukuoka, Japan). The placement order was repeated until the end of fertilization assessment. On day 3, embryos were scored using Puissant’s criterion [22]. Embryos were transferred, cryopreserved, or cultured to the blastocyst stage according to clinical indications and the willingness of the patient. During cycles of blastocyst culture, embryos were transferred to an incubator (K-MINC-1000, Cook Australia Pty. Ltd. Brisbane, Queensland, Australia) with a 5% oxygen concentration and extended cultured to days 5 to 7. Blastocysts were graded according to the Gardner and Schoolcraft system [23]. Blastocysts with a score ≥ 4BC or 4CB on day 5 were considered suitable for transfer, and the remaining blastocysts (≥ 4BC or 4CB) on days 5 to 7 were cryopreserved.
Outcome measures
All clinical outcomes were defined according to the reported guidelines [24]. The primary outcome was the CLBR. Other outcomes included the proportion of high-quality embryos (defined as a cell count ≥ 6 and a fragmentation rate < 20%), slow-developed embryos (defined as < 6 cells), fragmented embryos (with a fragmentation rate ≥ 20%) on day 3, cumulative clinical pregnancy, good birth outcomes (defined as a live birth at ≥ 37 weeks of gestation with a birthweight of 2500–4000 g and without a major congenital anomaly), and features of live births, including the rates of neonatal malformation, gestational hypertension, gestational diabetes, and preterm birth. In addition, gestational age, the interval between oocyte retrieval and live birth, birthweight, and Z score (defined as follows: [infant birthweight − mean birth weight at the same gestational age for the same sex in the reference population]/standard deviation [SD] in the reference population) [25] of live births were considered. For perinatal outcomes analyses, only the first of multiple live births during an oocyte retrieval cycle was included.
Statistical analysis
Continuous variables were reported as the mean ± standard deviation. A two-sample t-test was used for normally distributed values, and the Wilcoxon rank sum test was used for skewed data. Categorical variables were summarized as frequencies and percentages and analyzed using Pearson’s chi-square or Fisher’s exact test, as appropriate.
The baseline between the atmospheric and low oxygen groups was balanced using propensity score matching (PSM) [26]. Demographic and clinical factors that might confound the outcomes were taken into account, including female age, anti-Mullerian hormone (AMH) level, and body mass index (BMI) at the start of the cycle, years of infertility, diagnosis of polycystic ovary syndrome (PCOS), intrauterine adhesion and endometriosis, endometrial thickness, ovarian stimulation protocol, days of ovarian stimulation, total dosage of gonadotropins, number of oocytes retrieved, normal fertilization rate, male age, current male smoking status, sperm concentration, and sperm motility. A 1:1 nearest-neighbor caliper matching method was used to match the data between the 20% oxygen concentration group and 5% oxygen concentration group, and a caliper (0.01) of 0.2 of the standard deviation of the logit of the propensity score (0.05) was used [27, 28]. A standardized mean difference (SMD) of characteristics distribution of < 0.1 was considered indicative of a negligible difference between groups in the mean or prevalence of a covariate [29].
Sensitivity analyses of several important outcomes following the first transfer cycles between the groups assessed the robustness of the findings. The adjusted odds ratios (ORs) of outcomes analyzed with the full PSM data set were estimated using univariate logistic regression. Pre-specified subgroup analyses were conducted according to the age at the start of the cycle (strata: < 35 years, 35–37 years, 38–40 years, > 40 years). During the analysis of features of live births based on the live birth data set and subgroup analyses using subsets of each age stratum, multiple logistic regression, multiple linear regression, and linear mixed effect models were used as appropriate to adjust the matching covariates as well as the type of cycle (fresh or frozen), type of transferred embryos (cleavage or blastocyst), and number of embryos transferred. The estimated OR and β were reported with the 95% confidence interval (CI) and two-sided P value, with values < 0.05 considered significant. All analyses were performed using R software (R version 4.2.2).
Results
Our final analytic cohort included 31,566 women before the PSM according to the inclusion and exclusion criteria, as well as all fresh and thawed embryo transfer cycles associated with these retrieval cycles (Fig. (Fig.1).1). The 5% oxygen group had higher percentages of PCOS patients and male smokers than those in the 20% oxygen group. After 1:1 nearest-neighbor PSM matching, the final sample size of each group was 13,598, and the matching process resulted in a good balance of all covariates (SMD < 0.1) (Table (Table11).
Flow chart of patients involved in the analysis
Table 1
Demographic and clinical characteristics between unmatched and propensity score matched patients
| Unmatched | Matched | |||||
|---|---|---|---|---|---|---|
| 20% oxygen group n=14,288 | 5% oxygen group n=17,278 | SMD | 20% oxygen group n=13,598 | 5% oxygen group n=13,598 | SMD | |
| Female age at cycle started, mean (SD), years | 31.59 (±5.07) | 31.43 (±4.93) | 0.032 | 31.57 (±5.04) | 31.57 (±4.96) | <0.001 |
| Female BMI, mean (SD), kg/m2 | 22.02 (±2.72) | 21.97 (±2.64) | 0.020 | 21.97 (±2.70) | 22.00 (±2.63) | 0.009 |
| Female AMH, mean (SD), ng/ml | 5.32 (±4.52) | 5.44 (±4.59) | 0.026 | 5.24 (±4.44) | 5.31 (±4.44) | 0.016 |
| Infertility years, mean (SD), years | 3.95 (±3.11) | 3.78 (±2.90) | 0.056 | 3.84 (±2.98) | 3.89 (±3.01) | 0.015 |
| Infertility type (%) | 0.069 | 0.014 | ||||
| 4816 (33.7) | 6394 (37.0) | 4721 (34.7) | 4628 (34.0) | |||
| 9472 (66.3) | 10,884 (63.0) | 8877 (65.3) | 8970 (66.0) | |||
| PCOS (%) | 0.152 | 0.028 | ||||
| 1658 (11.6) | 2922 (16.9) | 1640 (12.1) | 1768 (13.0) | |||
| 12,630 (88.4) | 14,356 (83.1) | 11,958 (87.9) | 11,830 (87.0) | |||
| Intrauterine adhesion (%) | 0.076 | 0.002 | ||||
| 1273 (8.9) | 1935 (11.2) | 1259 (9.3) | 1268 (9.3) | |||
| 13,015 (91.1) | 15,343 (88.8) | 12,339 (90.7) | 12,330 (90.7) | |||
| Endometriosis (%) | 0.003 | 0.008 | ||||
| 1372 (9.6) | 1642 (9.5) | 1325 (9.7) | 1294 (9.5) | |||
| 12,916 (90.4) | 15,636 (90.5) | 12,273 (90.3) | 12,304 (90.5) | |||
| Ovarian stimulation protocol (%) | 0.024 | 0.010 | ||||
| 11,830 (82.8) | 14,448 (83.6) | 11,273 (82.9) | 11,288 (83.0) | |||
| 1847 (12.9) | 2157 (12.5) | 1765 (13.0) | 1734 (12.8) | |||
| 302 (2.1) | 335 (1.9) | 278 (2.0) | 292 (2.1) | |||
| 309 (2.2) | 338 (2.0) | 282 (2.1) | 284 (2.1) | |||
| Days of ovarian stimulation, mean (SD), days | 10.65 (±2.51) | 10.67 (±2.29) | 0.011 | 10.66 (±2.48) | 10.66 (±2.30) | <0.001 |
| Total dosage of gonadotropins, mean (SD), IU | 2290.70 (±1004.20) | 2304.49 (±992.95) | 0.014 | 2309.43 (±1004.45) | 2299.10 (±987.53) | 0.010 |
| Number of oocytes retrieved, mean (SD) | 11.09 (±6.11) | 11.21 (±5.91) | 0.021 | 11.08 (±6.09) | 11.13 (±5.91) | 0.008 |
| Normal fertilization rate, mean (SD) | 0.71 (±0.20) | 0.71 (±0.20) | 0.015 | 0.71 (±0.20) | 0.71 (±0.20) | 0.003 |
| Endometrial thickness, mean (SD), mm | 12.43 (±2.36) | 12.56 (±2.47) | 0.055 | 12.51 (±2.31) | 12.47 (±2.30) | 0.016 |
| Male age at cycle started, mean (SD), years | 33.70 (±5.92) | 33.49 (±5.74) | 0.035 | 33.65 (±5.91) | 33.66 (±5.77) | <0.001 |
| Male current smoking (%) | 0.111 | <0.001 | ||||
| 4280 (30.0) | 6076 (35.2) | 4248 (31.2) | 4251 (31.3) | |||
| 10,008 (70.0) | 11,202 (64.8) | 9350 (68.8) | 9347 (68.7) | |||
| Sperm concentration, mean (SD), million/mL | 56.42 (±18.79) | 56.80 (±19.94) | 0.020 | 56.55 (±18.04) | 56.42 (±17.54) | 0.008 |
| Sperm motility, mean (SD), % | 35.35 (±10.85) | 35.13 (±10.02) | 0.022 | 35.19 (±7.06) | 35.14 (±7.02) | 0.008 |
Data are presented as mean (± standard deviation) or number (%)
AMH antimullerian hormone, BMI body mass index, FSH follicle-stimulating hormone, PCOS polycystic ovary syndrome, SD standard deviation, SMD standardized mean difference
The embryo characteristics and outcomes of the 20% and 5% oxygen groups are presented in Table Table2.2. The rate of high-quality embryos in the 20% oxygen group was significantly lower than that in the 5% oxygen group (0.49 ± 0.33 vs. 0.51 ± 0.33; adjusted β = −0.03; 95% CI, −0.03 to −0.02), whereas the rate of slow-developing embryos was significantly higher (0.56 ± 0.31 vs. 0.52 ± 0.31; adjusted β = 0.04; 95% CI, 0.03–0.05). The rate of fragmented embryos did not significantly differ between the two groups (0.13 ± 0.22 vs. 0.13 ± 0.21; adjusted β = −0.002; 95% CI, −0.007 to 0.003). The 5% oxygen group had a significantly higher percentage of fresh embryo transfer cycles and a lower percentage of blastocyst transfer cycles than those in the 20% oxygen group. The percentage of non-transferable embryo cycles in the 20% oxygen group was higher than that in the 5% oxygen group (7.4% vs. 6.0%; adjusted OR = 1.26; 95% CI, 1.15–1.39).
Table 2
IVF outcomes of the cumulative transfer cycles between 20 and 5% oxygen groups
| 20% oxygen group | 5% oxygen group | Adjusted OR/β (95% CI) | P-value | |
|---|---|---|---|---|
| Features of embryo quality | ||||
| 8.41 (±5.00) | 8.49 (±4.89) | −0.08 (−0.20 to 0.03) | 0.168 | |
| 4.31 (±4.00) | 4.52 (±3.95) | −0.20 (−0.30 to −0.11) | <0.001 | |
| 0.49 (± 0.33) | 0.51 (± 0.33) | −0.03 (−0.03 to −0.02) | <0.001 | |
| 0.56 (± 0.31) | 0.52 (± 0.31) | 0.04 (0.03 to 0.05) | <0.001 | |
| 0.13 (± 0.22) | 0.13 (±0.21) | −0.002 (−0.007 to 0.003) | 0.429 | |
| Features of embryo transfer§ | ||||
| 15,591 | 15,735 | |||
| 63.7% (9933/15,591) | 66.0% (10,381/15,735) | 0.91 (0.86 to 0.95) | <0.001 | |
| 34.2% (5335/15,591) | 32.2% (5071/15,735) | 1.10 (1.04 to 1.15) | <0.001 | |
| 22.0% (3431/15,591) | 21.2% (3336/15,735) | 1.05 (0.99 to 1.12) | 0.087 | |
| Times of embryo transferred | 0.82 (0.75 to 0.89) | <0.001 | ||
| 10.1% (1370/13,598) | 8.4% (1144/13,598) | |||
| 69.0% (9378/13,598) | 71.0% (9655/13,598) | |||
| 17.7% (2412/13,598) | 17.4% (2372/13,598) | |||
| 3.2% (438/13,598) | 3.1% (427/13,598) | |||
| Rate of no transferable embryo cycle | 7.4% (1010/13,598) | 6.0% (814/13,598) | 1.26 (1.15 to 1.39) | <0.001 |
Data are presented as mean (± standard deviation) or percentage (event number/ total number)
§In the analytical process of features of embryo transfer based on cumulative live birth cycles data set, mixed effect logistic regression models was used to consider cluster effects for multiple transfer cycles from the same patient
The pregnancy and perinatal outcomes of the two groups are listed in Table Table3.3. The CLBR was significantly lower in the 20% oxygen group than in the 5% oxygen group (58.6% vs. 62.4%; adjusted OR = 0.85; 95% CI, 0.81–0.90). The neonatal malformation rate, gestational hypertension rate, gestational diabetes rate, preterm birth rate, and duration of pregnancy did not significantly differ between the two groups. There were 5602 and 5907 singleton live births in the 20% and 5% oxygen groups, respectively. The birthweight and Z score were significantly higher in the 20% oxygen group than those in the 5% oxygen group (birthweight: 3.30 ± 0.50 vs. 3.28 ± 0.48; adjusted β = 0.022; 95% CI, 0.004–0.040; Z score: 0.26 ± 1.04 vs. 0.22 ± 1.01; adjusted β = 0.037; 95% CI, 0.001–0.074). Regarding the multiple live births, birthweight did not significantly differ between the two groups. The cumulative rate of good birth outcomes was significantly lower in the 20% oxygen group than in the 5% oxygen group (34.5% vs. 36.6%; adjusted OR = 0.91; 95% CI, 0.87–0.96).
Table 3
Pregnancy and perinatal outcomes of the cumulative transfer cycles between 20 and 5% oxygen groups
| 20% oxygen group | 5% oxygen group | Adjusted OR/β (95% CI) | P-value | |
|---|---|---|---|---|
| No. of oocyte retrieval cycles | 13,598 | 13,598 | ||
| Cumulative live birth rate☩ | 58.6% (7970/13,598) | 62.4% (8483/13,598) | 0.85 (0.81–0.90) | <0.001 |
| 41.2% (5602/13,598) | 43.4% (5907/13,598) | 0.91 (0.87–0.96) | <0.001 | |
| 17.4% (2368/13,598) | 18.9% (2576/13,598) | 0.90 (0.85–0.96) | 0.001 | |
| Cumulative clinical pregnancy rate☩ | 67.3% (9147/13,598) | 71.0% (9655/13,598) | 0.84 (0.80–0.88) | <0.001 |
| Features of live births† | ||||
| 7970 | 8483 | |||
| 2.1% (166/7970) | 1.9% (160/8483) | 1.10 (0.89–1.37) | 0.342 | |
| 4.9% (387/7970) | 4.9% (419/8483) | 0.98 (0.85–1.13) | 0.801 | |
| 15.6% (1244/7970) | 16.1% (1366/8483) | 0.97 (0.89–1.06) | 0.523 | |
| 20.2% (1607/7960) | 21.1% (1788/8464) | 0.96 (0.88–1.04) | 0.351 | |
| 38.14 (±2.13) | 38.07 (±2.12) | 0.05 (−0.01–0.10) | 0.121 | |
| 306.30 (±120.15) | 300.70 (±110.69) | −0.04 (−2.66–2.58) | 0.975 | |
| 5602 | 5907 | |||
| 3.30 (±0.50) | 3.28 (±0.48) | 0.022 (0.004–0.040) | 0.016 | |
| 0.26 (±1.04) | 0.22 (±1.01) | 0.037 (0.001–0.074) | 0.046 | |
| 4749 | 5166 | |||
| 2.46 (0.47) | 2.46 (0.47) | 0.003 (−0.021–0.027) | 0.826 | |
| Cumulative good birth outcome rate* | 34.5% (4674/13,545) | 36.6% (4943/13,519) | 0.91 (0.87–0.96) | <0.001 |
Data are presented as mean (± standard deviation) or percentage (event number/ total number)
CI confidence interval, OR odds ratio
☩Twice or more live births/ clinical pregnancies under the same oocyte retrieval cycle are counted as one when the cumulative rates were calculated, the numbers of twice live births/clinical pregnancies in the 20% and 5% oxygen groups were 160 and 129, 650 and 598, respectively
†Only the first of multiple times of live births in an oocyte retrieval cycle was taken for the analysis. In the analytical process of features of live births, multiple logistics regression and multiple linear regression were used as appropriate to adjust the matching covariates, as well as the type of cycle (fresh or frozen), the type of transferred embryos (cleavage or blastocyst) and the number of embryos transferred
The association between birth weight in multiple fetuses and oxygen concentration in embryo culture was evaluated using a linear mixed effect model to consider cluster effects for multiplets
*The missing conditions of these indicators in the 20% and 5% oxygen groups are as follows: 10/7970 vs. 19/8483, 35/7970 vs.51/8483, 35/7970 vs. 52/8483, 42/5602 vs. 74/5907, 65/5602 vs. 97/5907, 86/4749 vs. 84/5166, 53/13598 vs. 79/13598
The clinical and perinatal outcomes of the first transfer cycle of the 20% and 5% oxygen groups are presented in the Supplementary Table 1. Live birth and clinical pregnancy rates were significantly lower in the 20% oxygen group than those in the 5% oxygen group (live birth: 53.1% vs. 56.0%; adjusted OR = 0.89; 95% CI, 0.85–0.94; clinical pregnancy: 63.6% vs. 66.0%; adjusted OR = 0.90; 95% CI, 0.86–0.95). The pregnancy loss rate was significantly higher in the 20% oxygen group than that in the 5% oxygen group (16.5% vs. 15.2%; adjusted OR = 1.10; 95% CI, 1.01–1.20). The neonatal malformation rate, gestational hypertension rate, gestational diabetes rate, preterm birth rate, and duration of pregnancy did not significantly differ between the two groups. Regarding singleton live births, the birthweight and Z score were significantly higher in the 20% oxygen group than those in the 5% oxygen group (birthweight: 3.29 ± 0.49 vs. 3.26 ± 0.48; adjusted β = 0.03; 95% CI, 0.01–0.05; Z score: 0.23 ± 1.03 vs. 0.18 ± 1.00; adjusted β = 0.05; 95% CI, 0.01–0.09).
The results of the age stratification analysis of the influence of oxygen concentrations on pregnancy outcomes are displayed in Fig. Fig.2.2. The CLBR, cumulative clinical pregnancy rate, and cumulative good outcome rate were higher in the 5% oxygen group than in the 20% oxygen group in all four age subgroups, and the proportion of non-transferable embryo cycles was lower in the 5% oxygen group. All these differences were significant in the subgroups of women aged < 35 and > 40 years.
Cumulative pregnancy outcomes of the 20% and 5% oxygen groups and age subgroups
Discussion
Our results indicated a significantly higher rate of high-quality embryos and a lower proportion of embryos with slow development on day 3 when embryos were cultured on days 1 to 3 at the 5% oxygen concentration compared with the embryos cultured at the 20% oxygen concentration. The cumulative clinical pregnancy and live birth rates in the 5% oxygen group were significantly higher than those in the 20% oxygen group. The birthweight and Z score of singletons in the 5% oxygen group were significantly lower than those in the 20% oxygen group.
In our study, a 5% oxygen concentration effectively improved embryo quality on day 3, which is consistent with previous research results [9, 30, 31]. Our data indicated that the 20% oxygen group had a higher proportion of embryos with fewer than six cells on day 3, confirming that the atmospheric oxygen concentration slowed cleavage embryonic development [32, 33]. The atmospheric oxygen concentration increased ROS production in embryos, leading to DNA damage. DNA damage repair genes are expressed during early embryonic development, and DNA repair depends on maternal mRNA and protein levels [34]. Therefore, the slowing of early embryonic development may be attributed to the prolonged time to repair DNA damage during cell cycle checkpoints [35]. There are some possible reasons for embryo fragmentation, such as maternal pathological factors, age, sperm quality, culture conditions, and chromosome abnormalities, but the exact mechanism of action is not yet understood [36]. Bedaiwy et al. reported no apparent relationship between day 1 ROS levels in culture media and embryo fragmentation in conventional IVF cycles [37]. Our results indicated that the oxygen concentration was not directly related to embryo fragmentation. Further research is needed to determine factors that increase the rate of embryo fragmentation.
Reduced embryo quality in the 20% oxygen group resulted in a lower cumulative clinical pregnancy rate and CLBR in our study. To explore whether the impact of the oxygen concentration on embryo quality was correlated to oocyte quality, we stratified the patients into four groups based on female age and conducted subgroup analyses. We observed that, regardless of age, the CLBR, pregnancy rate, and good birth outcomes of the 5% oxygen concentration group were higher than those of the 20% group, indicating that embryo development was damaged by high oxygen stress despite oocyte quality.
We observed significant differences between the two oxygen concentration groups in terms of singleton birthweights and Z scores of the cumulative live birth. We further performed sensitivity analyses of the first transfer cycles that transferred the best-quality embryos to verify the difference and observed that birthweight in the 20% oxygen group was also significantly higher than that in the 5% oxygen group. Brown et al. reported that the birthweight of calf embryos cultured under the 5% oxygen concentration was smaller than that of embryos cultured under the 20% oxygen concentration. The surface area of bovine embryo cotyledons cultured at the 20% oxygen concentration was increased, which could increase the area of nutrient exchange, accelerate embryonic development, and lead to fetal weight gain [38]. Our results suggested that exposing preimplantation embryos to an atmospheric oxygen concentration could result in long-term damage. It is noteworthy that while our results revealed that the birthweight of the 20% oxygen group was statistically higher than that of the 5% oxygen group, the difference was small (3300 g vs. 3280 g); the clinical significance of this difference should be interpreted carefully.
The strengths of this study include the use of PSM to balance the baseline characteristics between the two groups. We refined the embryo quality assessment indicators to explore the relationship between the oxygen concentration and embryo fragmentation, and we conducted an age stratification analysis to explore whether embryos from young patients can resist the damage of atmospheric oxygen concentration, thus providing more detailed reference information for relevant mechanism research. The limitations included the retrospective design. It is possible that these patients were characterized by unobservable differences that we cannot control. We performed follow-up for 1 to 5 years. At the end of the follow-up period, some patients had surplus frozen embryos that were preserved, which may bias the cumulative pregnancy rate. Furthermore, although our large sample size could improve the accuracy of statistical inference, it also increased the sensitivity of statistical tests, making it easier to detect differences or associations between groups, improving the statistical significance of the study. Therefore, the findings should be interpreted with caution in combination with clinical practice.
Conclusions
For embryos cultured from days 1 to 3 under the 5% oxygen concentration, the high-quality embryo rate and CLBR were significantly higher than those of embryos cultured under the 20% oxygen concentration, and the birthweight of singleton live births was lower than that of the 20% oxygen group. Our results substantiated that a culture environment with a low oxygen concentration was more suitable for embryo culture.
Supplementary information
Table S1. Clinical and perinatal outcomes of the first transfer cycle between 20% and 5% oxygen groups. (XLSX 12 kb)
Acknowledgements
We are grateful to all staff members of the IVF group at the Reproductive and Genetic Hospital of CITIC-Xiangya.
Author contribution
LC and SM were in charge of the conception, design, and drafting of the article. MX, FG, and CL contributed to data acquisition. SZ and GL were in charge of advice on the experimental design and revising the results. All authors contributed to the article and approved the submitted version.
Data availability
The datasets used and/or analyzed during the current study are available from the corresponding author upon reasonable request.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Longbin Chen and Shujuan Ma have contributed equally to this work and share first authorship.
Contributor Information
Shuoping Zhang, Email: moc.liamtoh@gnapiabiab.
Ge Lin, Email: nc.ude.usc@63fggnil.
References
Articles from Journal of Assisted Reproduction and Genetics are provided here courtesy of Springer Science+Business Media, LLC
Citations & impact
Impact metrics
Article citations
Oxidative Stress in Pregnancy.
Biomolecules, 13(12):1768, 09 Dec 2023
Cited by: 14 articles | PMID: 38136639 | PMCID: PMC10741771
Review
Free full text in Europe PMC
Similar Articles
To arrive at the top five similar articles we use a word-weighted algorithm to compare words from the Title and Abstract of each citation.
Reduced oxygen concentration during human IVF culture improves embryo utilization and cumulative pregnancy rates per cycle.
Hum Reprod Open, 2020(1):hoz036, 22 Jan 2020
Cited by: 24 articles | PMID: 31989044 | PMCID: PMC6975868
Impact of thyroid autoimmunity on the cumulative live birth rates after IVF/ICSI treatment cycles.
BMC Pregnancy Childbirth, 24(1):230, 02 Apr 2024
Cited by: 0 articles | PMID: 38566020 | PMCID: PMC10985846
The Livebirth Rate Per In Vitro Fertilization Cycle Is Higher Than The Cumulative Live Birth Rates of Intrauterine Insemination for Patients of Poseidon Group 3 With Unexplained Infertility.
Front Endocrinol (Lausanne), 12:768975, 01 Dec 2021
Cited by: 2 articles | PMID: 34925238 | PMCID: PMC8671868
Pregnancy and birth outcomes in couples with infertility with and without assisted reproductive technology: with an emphasis on US population-based studies.
Am J Obstet Gynecol, 217(3):270-281, 18 Mar 2017
Cited by: 72 articles | PMID: 28322775 | PMCID: PMC9761478
Review Free full text in Europe PMC