Retrospective cohort study on cumulative live birth rate of in vitro fertilization/intracytoplasmic sperm injection after multiple complete cycles in patients with diminished ovarian reserve.

We performed a retrospective cohort study including female patients undergoing IVF/ICSI cycle during the period from January 2016 to December 2020 at the Reproductive and Genetic Medical Center of Peking University First Hospital. All patients were followed up for 3 years until either discontinuation of treatment or delivery of one or more live infants. All data for fresh cycles and subsequent frozen-thawed cycles were analyzed. A complete cycle is defined as all transferable embryos being used after one oocyte retrieval cycle. All patients who meet the DOR diagnostic criteria but do not meet the exclusion criteria were all included in the study. No patient underwent preimplantation genetic test. Ethical approval was obtained from the Peking University First Hospital Ethics Committee. In this study, DOR was defined as serum anti-Mullerian hormone(AMH) level less than 1.1 ng/mL and bilateral antral follicle count(AFC) less than or equal to 6 at the first controlled ovarian hyperstimulation cycle. Exclusion criteria were as follows: Patients with submucosal fibroids, hydrosalpinx, uterine malformations, endometrial lesions, ovarian cancers, chromosome abnormality. Patients who have frozen embryos that have not been transplanted but have not achieved live birth. Patients with submucosal fibroids, hydrosalpinx, uterine malformations, endometrial lesions, ovarian cancers, chromosome abnormality. Patients who have frozen embryos that have not been transplanted but have not achieved live birth. Various protocols were used in ovulation induction according to the bilateral antral follicle counts and hormone levels of the certain cycles, such us gonadotropin-releasing hormone antagonist protocols, minimal stimulation protocols, progestin-primed ovarian stimulation protocols and pituitary down-regulation protocols (long protocols). The choices of fresh embryo transfers or frozen embryo transfers were based on the ovulation induction protocols, the number of available embryos, the endometrial conditions and the patient’s preferences. The numbers of transferred embryos were 1–3 per cycles. The luteal phases were supported by intramuscular progesterone, vaginal micronized progesterone or oral dydrogesterone. According to the BMI classification recommended for Chinese population, BMI is categorized into the following four groups, < 18.5 kg/m 2 (underweight), 18.5 to 23.9 kg/m 2 (normal weight), 24.0 to 27.9 kg/m 2 (overweight) and ≥ 28.0 kg/m 2 (obesity) [ 1 ]. Live birth was defined as the birth of one or more live infants with a gestation of at least 28 weeks after fresh or freeze–thaw embryo transfers. The details about the embryo rating, definition of high-quality embryos and low-quality embryos have been described in our previous studies [ 2 ]. The conservative estimate of CLBR and the optimal estimate of CLBR were calculated as outcome variables. A conservative estimate of CLBR is based on the assumption that patients who do not return for a subsequent cycle will not have a live birth in further treatment. The optimal estimate of CLBR is based on the assumption that patients who do not return for a subsequent cycle have the same live birth rates as those who return. The statistical analyses were performed with the IBM SPSS ver. 27.0. Categorical data are expressed as frequencies and percentage, and differences between groups were assessed with Pearson's Chi-square test. Descriptive data with non-normally distribution are expressed as median (inter-quartile range, IQR) and were compared using Mann–Whitney U test. Statistical significance was set at P value < 0.05. Logistic regression analyses was used to evaluate the factors associated with live birth.

A total of 343 patients were included in this study, including 1065 cycles. 69 cycles had been cancelled before egg retrieval. Baseline characteristics of the patients are shown in Table 1 . Table 1 Baseline characteristics Variables Number of patients Proportion Age（years)  ≤30 43 12.5%  31-33 53 15.5%  34-36 64 18.7%  37-39 74 21.6%  40-42 56 16.3%  ≥43 53 15.5% BMI (kg/m 2 )  ＜18.5 28 8.1%  18.5-23.9 203 59.2%  24.0-27.9 86 25.1%  ≥28.0 26 7.6% Duration of infertility  ＜3 years 170 49.6%  ≥3 years 173 50.4% Previous history of clinical pregnancy  Yes 176 51.3%  No 167 48.7% Previous history of delivery（≥28 weeks)  Yes 54 15.7%  No 289 84.3% AMH（ng/ml)  ≤0.5 163 47.5%  0.5＜AMH＜1.1 180 52.5% AFC  0-3 158 46.1%  4-6 185 53.9% Cause of infertility other than DOR  Male factor 112 32.7%  Tubal factor 83 24.2%  Endometriosis 36 10.5% Baseline characteristics The stimulation protocols and dosages of gonadotropins for each protocols of the 996 egg retrieval cycles are shown in Table 2 . A total of 474 embryo transfer procedures were involved. The types of embryos transferred are shown in Table 3 . The qualities of embryos transferred are shown in Table 4 . Table 2 The stimulation protocols and dosages of gonadotropins of 996 oocyte retrieval cycles Number of cycles Proportion Total dosages of gonadotropins[IU，median ( P 25 , P 75 )] Mild stimulation 402 40.4% 1200 (900,1725) Gonadotropin-releasing hormone antagonist protocol 361 36.2% 2700 (2175,3375) Progestin primed ovarian stimulation protocol 99 9.9% 2475 (1800,3000) Long protocol 69 6.9% 4200 (3375,4875) Natural and modified natural cycle protocol 65 6.5% 150 (150,450) Total 996 100% Table 3 The types of embryos transferred Type Number of embryos Proportion Fresh cleavage-stage embryo 495 57.3% Fresh blastocyst embryo 1 0.1% Frozen cleavage-stage embryo 302 35.0% Frozen blastocyst embryo 66 7.6% Total 864 100% Table 4 The qualities of embryos transferred Type Number of embryos Good-quality cleavage-stage embryo 506 Low-quality cleavage-stage embryo 291 Good-quality blastocyst embryo 37 Low-quality blastocyst embryo 30 Total 864 The stimulation protocols and dosages of gonadotropins of 996 oocyte retrieval cycles The types of embryos transferred The qualities of embryos transferred The number of patients who obtained oocyte retrieval operations, the number of patients who obtained live births and the number of patients who failed to obtained live birth but refused to receive further treatment in each cycle were showed in Fig. 1 . Considering the small number of patients underwent 9 or more cycles of oocyte retrieval treatment, their information will be merged into one group. Fig. 1 The number of patients receiving treatment and the number of patients obtained live births The number of patients receiving treatment and the number of patients obtained live births The conservative estimate of CLBR and the optimal estimate of CLBR of the whole population and in different age group are shown in the Fig. 2 . The live birth rate of the first oocyte retrieval cycle of the whole study subjects was 9.9%. After one or more cycles of treatments, a total of 141 couples achieved live births during the study period, with a conservative estimate CLBR of 41.1%. Assuming that the final outcome of the patients who did not continue to receive treatment had the same live birth rate with those who continued treatment, the optimistic estimate of CLBR was 81.0%. After calculating by age group, it was found that the older the female patient was, the lower the conservative and optimistic estimates of the CLBRs were. For women aged < 35 and 35 ~ 39, after 6 oocyte retrieval cycles, the conservative/optimistic estimates of CLBRs reached 57.4/82.7% and 41.1/64.8%, respectively. For elderly women ≥ 40 years old, the conservative/optimistic estimates of the CLBRs 6 oocyte retrieval cycles were only 14.7/26.0%. Fig. 2 The conservative and optimal estimates of CLBRs. A . whole population; B .patients age ＜35 years old; C .patients age 35~39 years old; D .patients age ≥40 years old) The conservative and optimal estimates of CLBRs. A . whole population; B .patients age ＜35 years old; C .patients age 35~39 years old; D .patients age ≥40 years old) The baseline characteristics, as well as the details regarding oocyte retrieval and embryo transfer, for patients who achieved live birth versus those who did not during the study period are presented in the Table 5 . There are statistical differences in age, AMH level, AFC, the history of clinical pregnancy, the history of live birth, the number of oocytes retrieved per cycle, the number of embryo transfer procedure and the accumulative number of embryos transferred are different between the two groups. Table 5 Baseline characteristics and treatment indicators of the two groups Patients obtained live birth Patients not obtained live birth P Number of patient 141 202 Age[years，median ( P 25 , P 75 )] 34(31, 38) 39 (35, 43) ＜0.01 AMH[ng/ml,median ( P 25 , P 75 )] 0.61 (0.35,0.79) 0.47 (0.20, 0.70) ＜0.01 AFC[median ( P 25 , P 75 )] 4 (3, 5) 3 (2, 5) ＜0.01 BMI[kg/m 2 , median ( P 25 , P 75 )] 22.0(19.9,24.5) 22.4 (20.3, 25.1) 0.22 Infertility duration[year,median ( P 25 , P 75 )] 2 (1, 4.5) 3 (1, 5) 0.4 History of clinical pregnancy(%) 40.4 58.9 ＜0.01 History of live birth(%) 8.5 20.8 ＜0.01 Number of oocytes retrieved per cycle [median( P 25 , P 75 )] 3(2, 4) 2 (1, 3) ＜0.01 Number of embryo transfer procedures [median ( P 25 , P 75 )] 2 (1, 2) 1 (0, 2) ＜0.01 Accumulative number of embryos transferred [median ( P 25 , P 75 )] 3.0(2.0,4.0) 2.0(0, 3.0) ＜0.01 Baseline characteristics and treatment indicators of the two groups Logistic regression analysis was performed on 8 variables mentioned before which were statistical difference between the group according to univariate analysis. The multivariate analysis results showed that: age and the number of oocyte retrieved per cycle were significantly associated with live birth outcomes (Table 6 ). The higher the age is, the lower the live birth rate is. A higher number of oocyte retrieved per cycle indicates a higher rate of live birth. Table 6 The relevant factor of live birth by the logistic regression Factor Parameter（B） SE Wald P OR 95% CI Age −0.14 0.03 21.00 ＜0.01 0.87 0.82-0.92 Number of oocytes retrieved 0.25 0.09 8.08 ＜0.01 1.28 1.08 −1.52 The relevant factor of live birth by the logistic regression There are 195 patients(dropout group) who discontinued treatments after one or more unsuccessful cycles. A total of 141 patients achieved live births during the study period. There are 7 patients who continued treatment after 9 or more oocyte retrieval cycles although they did not achieved live births yet. The baseline characteristics, the numbers of oocytes retrieved per cycle of the patients who discontinued treatments and other 148 patients(non-dropout group) during the period are showed in the Table 7 . There are statistical differences in age, AMH level, AFC and the number of oocytes retrieved per cycle between the two groups. Table 7 Baseline characteristics and treatment indicator of the dropout group and the non-dropout group Dropout Group Non-dropout group P Number of patient 195 148 Age[years，median ( P 25 , P 75 )] 39 (34, 43) 35 (31, 38) ＜0.01 AMH[ng/ml,median ( P 25 , P 75 )] 0.48(0.20,0.71) 0.58 (0.32, 0.78) ＜0.01 AFC[median ( P 25 , P 75 )] 3 (2, 5) 4 (3, 5) ＜0.01 Number of oocyte retrieved per cycle [median ( P 25 , P 75 )] 2(1, 3) 3(2, 4) ＜0.01 Baseline characteristics and treatment indicator of the dropout group and the non-dropout group The patients in dropout group received 512 cycles of oocyte retrieval, of which egg was not retrieved in 88 cycles (17.2%), and there was no transplantable embryo in 196 cycles (38.3%). The patients in non-dropout group received 484 cycles of oocyte retrieval, of which egg was not retrieved in 43 cycles (8.9%), and there was no transplantable embryo in 92 cycles (19.0%). There were significant differences in the proportion of cycles without oocytes and cycles without transplantable embryos between the two groups ( P  < 0.01, P  < 0.01). Treatment outcomes between different levels of AMH levels and different age groups are showed in the Tables 8 and 9 . Table 8 Treatment outcomes between different levels of AMH levels AMH(ng/ml) ＜0.5 0.5 ≤ AMH＜1.1 P Number of patients 161 182 / Number of egg retrieval cycles 538 458 / Number of cycles cancelled before egg retrieval 45 24 / Number of cycles without egg (%) 105(19.5%) 26(5.7%) ＜0.01 Number of cycles without transplantable embryo (%) 197(36.6%) 91(19.9%) ＜0.01 LBR/ET 27.4% 31.4% 0.36 Table 9 Treatment outcomes of different age groups Age (years) ＜35 35-39 ≥40 P Number of patient 122 112 109 / Number of egg retrieval cycles 351 325 320 / Number of cycles cancelled before egg retrieval 24 21 24 / Number of cycles without egg (%) 43(12.3%) 33(10.2%) 55(17.2%) 0.03 Number of cycles without transplantable embryo (%) 106(30.2%) 66(20.3%) 116(36.3%) ＜0.01 LBR/ET 39.1% 29.4% 16.5% ＜0.01 Treatment outcomes between different levels of AMH levels Treatment outcomes of different age groups

Diminished ovarian reserve(DOR) refers to the decrease in the quantity and quality of oocytes in the ovary, which leads to a decline in female fertility. The proportion of patients with DOR function is increasing due to factors such as delayed childbirth and environmental pollution. DOR is one of the main causes for women receiving in vitro fertilization and embryo transfer(IVF-ET) treatments. DOR may cause poor fertility outcomes even with the help of assisted reproductive techniques (ART). As a main challenge in ART, most patients will undergo multiple cycles of IVF treatments. Patients should be informed of the ‌estimated number of treatment cycles‌ required to achieve pregnancy. Compared with the pregnancy outcome of single cycle, the cumulative live birth rates (CLBRs) based on multiple complete IVF/intracytoplasmic sperm injection(ICSI) cycles might be more suitable to evaluate the efficacy of ART in DOR patients. Although there were currently some reports on CLBRs of IVF/ICSI population, the data of the entire population is not applicable to the DOR patients. There were few studies focus on the CLBRs in DOR patients who undertook IVF/ICSI treatments after multiple complete cycles. Here we report a retrospective cohort study on the CLBR of IVF/ICSI after multiple complete cycles in patients with DOR.

This cohort study shows that patients with diminished ovarian reserves can benefit from multiple IVF treatments. As is well known that age is the most important factor affecting a woman’s chance to success during IVF treatment, our findings are in agreement with those studies that CLBRs decrease sharply with the increase of ages [ 3 ]. At least 45% of female DOR patients younger than 40 years old could obtain a live birth after multiple cycles IVF treatment and the live birth rate could reach nearly 60% in patients younger than 35 years old. But for the elder patients of 40 years or more, the conservative estimates of the CLBR after multiple cycles sharply declined to 16.5%. The real CLBR is between the conservative and optimal estimates of CLBRs. The conservative CLBRs of DOR patients  2% increase per additional oocyte retrieved cycles) which indicate that IVF treatments could reduce the gaps in the probability of pregnancy between the young DOR patients and other female patients undergoing IVF, which are in line with studies published by Khalife and Smith [ 4 , 5 ]. However, In women aged 40 or above, the conservative CLBR was significantly low with a plateau(≤ 2% increase in CLBR per additional oocyte retrieved cycles) after 4 cycles reaching 13.8%, which is in line with previous studies reporting that IVF treatments could not compensate for the decrease of fertility among women over 40 years old [ 6 ]. These results may help reproductive medicine physicians counseling their infertility female patients with DOR about the chances of successful live births with repeated IVF cycles. These findings are encouraging for young DOR patients insisting to extend IVF treatment up to 6 cycles, fulfilling more opportunity for embryo transfer which is closely related to live birth outcome. As for elderly patients aged 40 or above, more than four cycles of IVF/ICSI may have limited benefits. As there are few effective interventions exist to improve the ovarian reserve, elderly DOR patients should have detailed consultation about the low live birth rate of single cycle and multiple cycles before starting treatment. AMH and AFC are commonly used indicators to evaluate ovarian reserve function in daily clinical practice [ 7 ]. AMH is an objective measurement index and does not fluctuate with the menstrual cycle. The measurement of AFC is subject to certain conditions such us the experience of the measurer and the resolution of the ultrasound instrument. AMH and AFC are all closely related to the outcome of assisted pregnancy treatments. But for some patients, the results of AMH and AFC are inconsistent. We used the low AMH level as well as low AFC as indicators to screen a group of patients truly suffering from diminished ovarian reserve and try to figure out their pregnancy outcomes. The significance difference between the conservative estimates of CLBR and optimal estimates of CLBR is associated with the high dropout rate of the DOR patients. Compared to women with normal ovarian reverse, the numbers of oocytes retrieval and the numbers of available embryos were less in DOR patients. It may really take a long time for them to obtain a live birth from IVF treatments. The pregnancy and live birth rates were relatively low and the patients may discontinue the following IVF/ICSI treatments because of the poor prognosis. Owing to the difference of the medical environment between China and some other countries, patients have free choice of medical institutions to seek for help. Patients who do not have available embryos after IVF treatments are more likely to switch to another hospitals for treatments. The results of this study showed that patients who discontinued treatment after one or more unsuccessful cycles were older, had lower AMH and AFC. The patients who discontinued had fewer oocytes per cycle and higher proportion of cycles without oocytes and embryos. The other reasons for the patients’ discontinuance of further IVF/ICSI in this medical institution may be as follows: seeking treatment from another institution, economic factors or giving up further artificial assisted pregnancy treatment [ 8 ]. The high dropout rate was also mentioned in other study with the potential causes of physical factor, emotional factor, and financial strain [ 9 ]. This may be the reason for the gap between the conservative and optimal estimates of CLBRs. The main advantage of this study is the long follow-up period. Prolonged follow-up was performed for 3 years until a live birth was obtained or all available embryos were transferred. Patients of DOR have long treatment cycles. They may need to receive several times of controlled ovarian hyperstimulations and oocytes retrieval treatments to accumulate embryos. One cycle of treatment cannot represent their prognosis. The findings from this study on CLBRs after multiple cycles of IVF/ICSI treatments may be helpful to both infertile couples and clinicians. Repeated IVF/ICSI cycles are beneficial for DOR patients especially for younger patients. The limitation of this study was its retrospective design and the relatively small scale of patients. Actually, we have treated more DOR patients than the number in statistics. Currently, some of the patients still store spare frozen embryos in our embryo laboratory and we could not know the pregnancy outcome of the oocytes retrieval cycle until all available embryos used up. The pregnancy rate is relatively low when transplanting single or double embryos of non top quality under morphological evaluation. What’s more, the operation of embryo transfer also incurs a cost. If the couples made decision to choose another institution to search for treatment, they may abandoned the embryos stored in our hospital. As the result, the patients may still store some embryos of low qualities and they cannot be included in the statistics.

For female DOR patients younger than 40 years old, multiple cycles IVF/ICSI treatments can help to achieve satisfactory success rates. Physicians may encourage them to try more cycles of IVF/ICSI treatment to gain more chances of live birth. As for elderly patients aged 40 or above, more than four cycles of IVF/ICSI may have limited benefits. If they fail to achieve live birth after trying multiple cycles of IVF/ICSI, they may choose other treatment methods.