Elective and Onco-fertility Preservation: Factors Related to IVF Outcomes

A. Cobo; J. García-Velasco; J. Domingo; A. Pellicer; J. Remohí


Hum Reprod. 2018;33(12):2222-2231. 

In This Article


Age-related fertility decline was the reason for oocyte vitrification in the EFP group. The majority of patients in the Onco-FP group were diagnosed with breast cancer (64.6%), followed by women afflicted by Hodgkin's (11.6%) and non-Hodgkin's lymphoma (5.2%). A minority comprised patients diagnosed with other malignancies that affected ovary (2.9%), digestive system (2.7%), thyroid (1.5%), uterine cervix (1.0%), leukemia (1.0%), colon (0.8%), bone (0.7%), lung (0.2%) and endometrium (0.2%). The remaining 7.5% had other cancer types. Supplementary Figure S1 shows the percentages of FP procedures compared to the total vitrification procedures in our clinics, including oocyte vitrification in autologous cycles for reasons other than FP and the vitrification of donors' oocytes over a 10-year period (N = 57212 total oocyte vitrification procedures). In the last 2 years, the vitrification of oocytes due to FP represents 25% of all procedures. Increasing trends are notable for EFP, and have grown from ~4% in 2007 to 22% in the last 2 years. Conversely in the Onco-FP group, this trend has remained quite steady (1%–2%) over the years.

Supplementary Figure S1.

Percentages of fertility preservation (FP) procedures with respect to the total vitrification procedures including oocytes vitrification in autologous cycles for other reasons than FP and vitrification of donors' oocytes in ten years period. Onco-FP (blue), N = 1139; elective-FP (EFP) (red), N = 6465; Total FP (green), N = 57212.

Table I shows the baseline characteristics. The number of COS/vitrification cycles per patient was statistically significantly higher in the EFP group. The EFP patients were significantly older (mean age = 37.2 ± 4.9 year) than the Onco-FP patients (mean age = 32.3 ± 3.5 year) (P < 0.0001). As shown in Panel A of Supplementary Figure S2, the vast majority (N = 4288, 81.1%) of the EFP women had their oocytes vitrified after the age of 35 year, 72.6% when aged between 35–40 year and 15.8% when aged 40 year or older. On the contrary, the oocytes of most of the Onco-FP group women (N = 751 patients, 69.9%) were vitrified at ≤35 year of age (Panel B of Supplementary Figure SII).

Supplementary Figure S2.

Distribution of patients' age at vitrification in elective fertility preservation (EFP) and Onco-FP groups. Green: <30 year, Blue: 30–35 year, Red: 36–39 year, Gray: ≥ 40 year. Tables show data according to specific age groups.

The antagonist protocol was used more frequently in the EFP group than in the Onco-FP one (percentage of using the antagonist protocol in EFP vs. Onco-FP = 91.1% versus 24.1%, respectively) (P < 0.05). On the contrary, the antagonist plus letrozole protocol was more widely used in the Onco-FP group (72.8%) and was not used for EFP cycles. After comparing the intra-EFP and Onco-FP groups, the number of retrieved and vitrified oocytes per COS cycle was larger with the antagonist protocol (Supplementary Table SI). We also vitrified fewer oocytes when we used the antagonist plus letrozole protocol in cancer patients compared to the antagonist protocol in this group (8.6 ± 6.6 vs. 10.3 ± 7.5; P < 0.05). Antral follicular count (AFC) was lower in EFP patients (Table I). Despite the anti-Müllerian hormone (AMH) levels also being higher in the Onco-FP group, we found no statistical differences between groups (NS). Supplementary Table SII shows the AFC and AMH data according to age. As expected, the AFC and AMH values were statistically higher for the patients aged ≤35 year versus >35 year in both FP groups (P < 0.0001). When considering only young patients, AFC was higher in the ONCO-FP group (13.9 vs. 11.1; P < 0.0001). When we separately analyzed older patients, AFC and AMC were comparable between the EFP and Onco-FP groups (NS).

Stimulation length and the dose of other gonadotropins during COS gave no statistical differences, except for the FSH dose, which was higher in the EFP group (Table I). It took longer from the day of first consultation to initiate COS in the EFP group (mean days = 19.7 ± 6.1 vs. 9.2 ± 11.3; P < 0.05). The E2 level on the day of ovulation triggering was considerably lower in the Onco-FP patients (Table I). In all, 67544 oocytes (mean per cycle = 9.6 ± 8.4) were retrieved in EFP, and 13381 oocytes in the Onco EFP group (mean 11.4 ± 3.5) (P < 0.05). Likewise, the mean number of oocytes vitrified per cycle was statistically lower in the EFP group (7.3 ± 11.3 vs. 8.7 ± 2.1) (P < 0.05). However, when analyzed per patient, the numbers of retrieved and vitrified oocytes were comparable (NS) reflecting the larger number of cycles per patient in the EFP group (Table I).

Table II shows the IVF data and clinical outcome of the EFP (N = 641) and the Onco-FP patients (N = 80) who returned to attempt pregnancy with their vitrified oocytes. The returning rate was higher in EFP (12.1% vs. 7.4%) (P < 0.05). It took longer for cancer patients to return (mean storage time 4.1 ± 0.9 vs. 2.1 ± 1.6 years) (P < 0.05). The number of oocytes warmed per patient was higher in the EFP group (9.1 ± 3.8 vs. 7.5 ± 2.8; P < 0.05). A significantly higher proportion of warming cycles ended up in ET for cancer patients (72.5% in Onco-FP vs. 50.2% in EFP) (P < 0.05). The implantation rate was higher in the EFP patients (42.6% vs. 32.5%) (P < 0.05). Although the clinical and ongoing pregnancy rates were higher in the EFP group, no statistical differences were observed (NS) (Table II). A total of 115 (EFP) and 18 (Onco-FP) babies were born, and 159 (24.8%) and 21 (56.9%) patients had cryo-transfers of surplus embryos (P < 0.05) of which 47 and 7 more babies were born (Table II).

Table III shows the survival and cumulative clinical outcomes in both groups according to patients' age at vitrification. When calculated on a per cycle basis, it refers to the vitrification/warming cycles of oocytes. The CLBR was calculated per patient, excluding the pregnant women who had not reached their due date when the analysis was done (patients who got pregnant from August 2017 onward). Comparisons were made between the EFP and Onco-FP groups for those patients aged ≤35 and >35 year. Survival (91.4% vs. 81.2%), the cumulative clinical pregnancy rate (CCPR = 65.9% vs. 42.8%), the cumulative ongoing pregnancy rate (COPR = 57.7% vs. 35.7%) and the CLBR (68.8% vs. 42.1%) were statistically higher for the EFP patients aged ≤35 year than for the Onco-FP age-matching women (P < 0.05) (Table III). By contrast in the older group (patients >35 year at vitrification) results were comparable between the EFP and Onco-FP groups (NS). As expected, younger patients obtained higher outcomes than the older ones when making intra-FP group comparisons (P < 0.05).

In order to rule out the effects of the indication for FP (EFP and Onco-FP) and age on oocyte survival and the CLBR, a BLR analysis was performed. As shown by the OR, the indication per se had no effect on either survival (OR = 1.484 [95%CI = 0.876–2.252]) (P = 0.202) or the CLBR (OR = 1.275 [95%CI = 0.711–2.284]) (P = 0.414). However, as Supplementary Table SIIIA indicates, the effect of the reason for FP on survival was acknowledged when age, categorized as ≤35 year and ≥36 year, was included in the model (adj. OR = 1.968 [95%CI= 1.121–3.445]; P = 0.018). The effect of age per se on survival was also confirmed (adj. OR = 1.922 [95%CI = 1.274–2.900]; P = 0.025). Similarly, the effect of the indication for FP (adj.OR = 2.204 [95%CI = 1.162–4.183]; P = 0.016) and age (adj. OR = 3.106 [95%CI = 2.039–4.733]; P < 0.0001) on the CLBR was confirmed (Supplementary Table SIIIB).

The effect of age was confirmed when other subcategorisations (≤30; 31–35; 36–40 and >40 year) were made (P < 0.05) (Supplementary Table SIIIA–B). The BLR model, including age, indication for FP and the COS protocol, confirmed the effect of age at vitrification on both survival and the chance of a live birth (P < 0.05) (Supplementary Table SIVA–B). However, the effect of the indication for FP was not statistically significant in this model (NS) neither was the possible effect of the COS protocol (NS) as shown by the corresponding adj.OR and the 95%CI (Supplementary Table SIVA–B).

The cumulative probability of achieving at least one baby according to age at vitrification, indication for FP and utilized oocytes is shown in Figure 1A, which provides the Kaplan–Meier plotting for EFP (Panel A) and Onco-FP (Panel B). In the EFP group, overall comparisons indicated significantly higher outcomes for younger patients than for older ones (P < 0.0001). The pace at which the CLBR rose was higher for those patients aged ≤35 year. The curves for older patients reached the plateau earlier than those for young women. For example, the cumulative probability of live birth (CLBR) doubled when three additional oocytes were used in younger patients (CLBR = 15.8% (95%CI = 8.4–23.1) with five oocytes to 32.0% (95%CI = 22.1–41.9) with eight oocytes used), while the increase in success was less marked in older patients (CLBR = 5.9% (95%CI = 3.6–8.3) and 17.3% (95%CI = 13.3–21.3) with 5 and 8 oocytes used, respectively). For younger patients, a reasonable number of 10 or 15 oocytes yielded success rates of 42.8% (95%CI = 31.7–53.9) and 69.85% (95%CI = 57.4–82.2) for the CLBR. The plateau in the subgroup of young women (≤35 year) was reached with 24 oocytes, and also with a remarkably high success rate (94.4% [95%CI = 84.3–100.4] CLBR).

Figure 1.

Kaplan–Meier plotting of cumulative probability of live birth (CLBR) for elective fertility preservation (EFP) (A) and Onco-FP (B), according to the number of oocytes consumed and patients' age at vitrification (≤35 year. in green and >35 year. in blue). Overall comparisons (log-rank (Mantel–Cox), Breslow (generalized-Wilcoxon) and Tarone–Ware) for EFP = P < 0.0001. Log-rank (Mantel–Cox); P = 0.577; Breslow (generalized-Wilcoxon); P = 0.833; and Tarone–Ware; P = 0.703 for onco-FP group. Tables below the figures show CLBR and 95%CI according to the number of oocytes consumed in each case.

Contrarily to what we observed in the EFP patients, the curves in the Onco-FP group practically overlapped, most probably due to the small sample size of the patients who returned to attempt pregnancy (Figure 1B). Accordingly, no statistical differences in the CLBR were observed (Log-rank (Mantel–Cox), P = 0.577; Breslow (generalized-Wilcoxon), P = 0.833; and Tarone–Ware, P = 0.703).