Basic Characteristics of the PatientsGeneral Description of the Donor Population

In total, data from 399 donors were used in this study. Each donor underwent a minimum of two and a maximum of nine cycles. The overall mean age of the donors was 25.45 ± 4.074 years, ranging from 19 to 35 years, and there were 102 donors (25.56%) below the age of 23 years. A total of 1965 ETs were performed among the 399 donors.

Independent Median Test to Compare Laboratory Indicators and Pregnancy Outcomes Between COS Cycle Groups

The results presented in Table 1 show that the median maturation rate per donor per cycle (80, 80, 79.31, 80.77, 82.32, 83.33 for cycle group 1, 2, 3, 4, 5 and 6 respectively; p < 0.0001), total number of oocytes used per donor per cycle (30, 32, 35, 35, 38, 39 for cycle group 1, 2, 3, 4, 5 and 6 respectively; p < 0.0001), and total MII used per donor per cycle (23, 25, 27, 28, 30, 32 for cycle group 1, 2, 3, 4, 5 and 6 respectively; p < 0.0001) were significantly different between the COS cycle groups. The comparative analysis of mean rates pertaining to Maturation, Fertilization, VEC, and Blastocyst outcomes across OPU groups is additionally presented in Fig. 1.

Fig. 1

Maturation, Fertilization Rates, VEC and Blastocyst Rates Across OPU Groups. VEC: Number of blastocytes at grades equal or grater than 3BB. Blast Rate: Number of blastocytes at grades equal or grater than early blastocytes (grades 1 and 2). Maturation rate: Number of mature oocytes/total number of oocytes collected. Fertilization rate: Number of fertilized oocytes (2PN and 2polar body)/number of mature oocytes injected (M2)

When pairwise comparison was performed on maturation rate per donor per cycle to determine which COS cycle groups were responsible for the significant difference, we found that cycle group 6 was significantly different compared with cycle group 2 (p < 0.0001) and cycle group 3 (p = 0.005). Furthermore, the maturation rate per donor per cycle was higher in cycle group 6 (83.33%) than in both cycle group 2 (80%) and cycle group 3 (79.31%).

Regarding the total number of oocytes used per donor per cycle, compared with cycle group 1, significant differences were observed in cycle group 3 (p = 0.025), 4 (p = 0.001), 5 (p < 0.0001), and 6 (p < 0.0001). The total number of oocytes used per donor per cycle was lowest in cycle group 1 (30) compared with the other cycle groups (35 in cycle groups 3 and 4, 38 in cycle group 5, and 39 in cycle group 6). In addition, significant differences regarding the total number of oocytes used were observed between cycle groups 2 and 5 (32 [0–96] and 38 [11–106], respectively; p = 0.041), 2 and 6 (32 [0–96] and 39 [8–105], respectively; p < 0.0001), and 3 and 6 (35 [6–100] and 39 [8–105], respectively; p = 0.010). Therefore, as the number of repeated COS cycles increased, the total number of oocytes used per donor per cycle also increased significantly.

Regarding the total MII used per donor per cycle, compared with cycle group 1, significant differences were observed in cycle group 3 (p = 0.021), 4 (p = 0.014), 5 (p = 0.001), and 6 (p < 0.0001). The total MII used per donor per cycle was lowest in cycle group 1 (23) compared with the other groups (27, 28, 30, and 32 in cycle groups 3, 4, 5, and 6, respectively). In addition, significant differences regarding the total MII used were observed between cycle groups 2 and 5 (25 [0–83] and 30 [0–96], respectively; p = 0.020), 2 and 6 (25 [0–83] and 32 [0–88], respectively; p < 0.0001), and 3 and 6 (27 [8–32] and 32 [0–88], respectively; p = 0.003). Therefore, as the number of repeated COS cycles increased, the total number of MII used per donor per cycle also increased significantly.

Moreover, reproductive outcomes were examined concerning repeated donor COS cycles, revealing no statistically significant differences in relation to both pregnancy and live birth (Table 1).

Prior to employing the GLMM, Spearman's correlation coefficient analysis was conducted to elucidate the relationships between OPU Groups and donor age. Additionally, correlations were explored between OPU Groups and average values calculated at the donor level for several factors initially measured at the cycle level. These factors included fertilization rate, number of 2PN, total number of thawed embryos, total number of oocytes used, total M2 used, number of embryos transferred, number of embryos frozen, VEC rate, and the number of blastocyst embryos.

Furthermore, correlations were investigated between OPU Groups and average values calculated at the donor level for factors initially measured at the OPU level, such as the total number of oocytes, total M2, Abortus rate, and Maturation rate. These analyses were crucial for comprehensively understanding the interplay between OPU Groups and various reproductive parameters, shedding light on the nuanced effects of repeated COS cycles within the studied population.

The Spearman's correlation coefficient test revealed a non-significant correlation between donor age and OPU groups (p = 0.054). However, other parameters demonstrated statistically significant relationships with OPU groups at the donor level.

The average value of 2PN embryos at the donor level had a subtle but statistically significant positive correlation (r = 0.045, p = 0.048) with OPU groups. Similarly, the average value of thawed embryos (r = 0.113, p =  < 0.0001) and average number of frozen embryos (r = 0.062, p = 0.007) all exhibited positive correlations with OPU groups at the donor level.

Additionally, the average number of total oocytes (r = 0.155, p =  < 0.0001), average number of M2 (metaphase II) (r = 0.160, p =  < 0.0001), and average number of abortus rate (r = 0.074, p = 0.001) showed statistically significant positive correlations with OPU groups at the donor level.

In the realm of reproductive sciences, these findings highlight the intricate relationships between repeated OPU cycles and various reproductive parameters, offering insights into embryonic development, cryopreservation outcomes, and oocyte quantity.

Analyzing Effect of Repeated COS cycles on Each Dependent Variable Using GLMM Models

To assess donor performance, four separate models were initially computed using GLMM models, taking the dependent variable as i) maturation rate per donor per cycle, ii) fertilization rate per donor per cycle, iii) VEC rate per donor per cycle, and iv) blastulation rate per donor per cycle. The independent variable was the COS cycle group. In all four models, the reference group used in the independent variable COS cycle groups was cycle group 1, that is, the effect of a patient having repeated COS cycles on the outcome variable was compared to cycle group 1.

From the results of the GLMM models for maturation rate, fertilization rate, VEC rate, and blastulation rate per donor per cycle, the donors who underwent 2, 3, 4, or 5 COS cycles did not have significantly different maturation rates, with a mean rate of 77% (p = 0.520, p = 0.064, p = 0.055, and p = 0.904 for cycles 2, 3, 4, and 5, respectively, compared to cycle 1), nor did they have significantly different fertilization rates, with a mean rate of 72.1% (p = 0.154, p = 0.682, p = 0.142, and p = 0.283 for cycles 2, 3, 4, and 5, respectively, compared to cycle 1), VEC rates, with a mean rate of 56.3% (p = 0.140, p = 0.240, p = 0.209, and p = 0.155 for cycles 2, 3, 4, and 5, respectively, compared to cycle 1), or blastulation rates, with a mean rate of 57.5% per donor per cycle (p = 0.194, p = 0.254, p = 0.165, and p = 0.132 for cycles 2, 3, 4, and 5, respectively, compared to cycle 1). Significant differences in maturation rate (p < 0.0001), fertilization rate (p = 0.020), VEC rate (p = 0.021), and blastulation rate (p = 0.015) were observed only for donors who underwent six or more COS cycles compared to one cycle. Therefore, ovarian response measured in terms of VEC rate and blastulation rate was similar in patients undergoing two to five cycles of ovarian stimulation. However, when the patient underwent six or more cycles of ovarian stimulation, even when both the maturation rate per donor per cycle and fertilization rate per donor per cycle increased by 3.9% and 2.2%, respectively, the VEC rate per donor per cycle and blastulation rate per donor per cycle decreased by 4.5% and 4.7%, respectively.

To make further conclusions on donor performance, two additional GLMM models were computed using i) the VEC rate per donor per cycle (Table 2) and ii) the blastulation rate per donor per cycle (Table 3) as dependent variables. To identify the factors affecting the VEC and blastulation rates per donor per cycle, fertilization rate and maturation rate per donor per cycle were first grouped into two categories each. If the donor’s fertilization rate per cycle was below the average rate, the donor was assigned to group 1; if the donor’s fertilization rate per cycle was above the average rate, the donor was assigned to group 2. A similar process was performed for maturation rate per donor per cycle, which was treated as a categorical variable in the model. This enabled us to determine whether the VEC and blastulation rates per cycle were affected by the donor having below or above average rates of fertilization and maturation per cycle. Therefore, in such models and in COS cycle groups, donor age, fertilization rate, and maturation rate per donor per cycle, which were treated as categorical variables, were added to the model as independent variables. The analytical results derived from the two models suggest that factors such as donor age, fertilization rate, and maturation rate exhibit statistically non-significant associations with the VEC rate (p = 0.161, p = 0.439, p = 0.057, respectively) and the blastulation rate per donor per cycle (p = 0.805, p = 0.271, p = 0.219, respectively). Upon meticulous adjustment for the covariates of donor age, fertilization rate, and maturation rate per donor per cycle, our investigation demonstrates a noticeable attenuation in donor performance, as measured by the VEC rate per cycle (p = 0.034) and the blastulation rate per cycle (p = 0.024), as IVF cycles extend beyond the fifth cycle. In specific terms, the VEC rate per cycle exhibits a decrease of 4.2%, while the blastulation rate per cycle witnesses a reduction of 4.5%. These findings underscore the significance of accounting for the cumulative effect of IVF cycles on donor performance metrics.

Table 2 GLMM Model of VEC rate per donor per CYCLETable 3 GLMM Model of blastulation rate per donor per cycle