This study included 14 311 women, with 17 315 embryo transfer cycles and 6331 live births, and the overall live birth rate was 40.54% in the first cycle, rising by the fourth cycle up to 44.24% (Fig. 1 ). The characteristics are summarized in Table 1 . About 48.84% of the women were younger than 30 years of age, and 8.80% were older than 37 years of age. Most of the women (61.16%) were at normal BMI. Tubal factor (62.48%) was the main cause of infertility.
Figure 1. Flow chart over included cases and their course of treatments, women followed until their first live birth.
Characteristic Women (n=14 311) Age (year) Mean±SD 30.95±4.65 ≤30 6990 (48.84) 30–34 4245 (29.66) 34–37 1816 (12.69) >37 1260 (8.80) BMI (kg/m2) Mean±SD 22.7±3.31 <18.5 963 (6.73) 18.5–24 8753 (61.16) 24–28 3238 (22.63) ≥28 996 (6.96) Missing 361 (2.52) Duration of infertility (year) Mean±SD 4.04±3.13 <2 2693 (18.82) 2–3 2684 (18.75) 3–5 3911 (27.33) ≥5 4447 (31.07) Missing 576 (4.02) Type of infertility Primary infertility 8387 (58.61) Secondary infertility 5761 (40.26) Missing 163 (1.14) Infertility diagnosis* Male factor 3169 (22.14) Endometriosis 1118 (7.81) Ovulation disorder or PCOS 1559 (10.89) Tubal factor 8941 (62.48) Unexplained factor 475 (3.32) Insemination method IVF 10 325 (72.15) ICSI 3063 (21.40) IVF+ICSI 602 (4.21) Missing 321 (2.24) Endometrial thickness (mm) Mean±SD 10.39±2.36 <7 415 (2.90) ≥7 10 602 (74.08) Missing 3294 (23.02) *Multiple diagnoses were possible, so totals are greater than 100%; PCOS: polycystic ovarian syndrome; IVF: in vitro fertilization; ICSI: intracytoplasmic sperm injection.
Table 1. Baseline demographic and clinical characteristics of the study population
Live birth occurred in 36.56% of the cycles, and 44.24% of the women had a live birth. In the first cycle, of women who failed to get pregnant, 6021 (42.07%) quited the ART treatment and 2489 women (17.39%) went on with a second treatment cycle; in the second cycle, of women who were not pregnant, 1611 quitted and 419 continued; in the third cycle, just 96 chose a fourth cycle. The specific cumulative live birth rate of each cycle is detailed in Supplementary Table 1 (available online).
We found a progressive decline in both optimal and conservative estimates of the cumulative live birth rate with increasing female age (P<0.001) (Fig. 2A and B ). By the fourth transfer cycle, the conservative and optimal estimates of live birth rates were 52.95% and 77.30% in women younger than 30 years of age and 18.17% and 26.51% in those 37 years of age or older, respectively.
Fig. 2C and D show the optimal and conservative live birth rates at different measures of female BMI. The conservative and optimal estimates of live birth rates after three transfer cycles were 46.22% and 61.97% in the normal BMI group (18.5–24 kg/m2), and 33.84% and 53.90% in the obese group (≥28 kg/m2), respectively. The contents of Fig. 2C and D indicate that obese women have lower optimal and conservative live birth rates than other women (P<0.001). Interestingly, as showed inFig. 2C , there was no statistical significance in the optimal estimates between those groups whose BMI lower than 28 kg/m2 (P>0.05).
Fig. 3A and B show the correlations between duration of infertility and live birth rates. In the analysis of duration of infertility, there is a significant decrease in both optimal and conservative estimates of the cumulative live birth rate for those women with the duration of infertility exceeding 5 years. For instance, compared to those groups whose duration of infertility below 5 years, the conservative estimates declined by 7.96%, 8.19%, and 7.13%, while the optimal estimates declined by 5.63%, 13.51%, and 6.39%.
Fig. 3C and D show that, from the first cycle to the fourth, the live birth rate of women with endometrium thickness greater than 7 mm is about 1.5 times than those with endometrial thickness less than 7 mm (conservative estimate 44.70% vs. 32.05%, while optimal estimate 63.15% vs. 46.18%). By the fourth transfer cycle, the conservative and optimal estimates of live birth rates were both above 45% for women with endometrium thickness greater than 7 mm.
Fig. 4A and B show that the CLBRs of women with primary infertility was slightly higher than those with secondary infertility (4.14% and 7.80% respectively for conservative and optimal estimate) but showing no significant statistical difference. Apart from unexplained factor, women whose infertility is caused by male factor had the highest CLBRs of all types after the third complete cycle (the conservative and optimal estimates, 61.79% and 75.55%, repectively), while those with a diagnosis of tubal factor infertility had the lowest ones (42.10% and 60.54%, respectively) (Fig. 4C and D ).
Fig. 5 illustrates the conservative and optimal CLBRs stratified by different treatments in the first complete cycle. Among all groups, IVF plus ICSI had higher conservative and optimal estimates of CLBRs (49.00% and 72.26%). However, we can't investigate the real disparity among those groups because most of women had choosen the IVF and ICSI caused big population number difference between different methods.
Characteristics of the study population
Live birth rates and female age
Live birth rates and female BMI
Live birth rates and the duration of infertility
Live birth rates and endometrial thickness
Live birth rates and type of infertility and infertility diagnosis
Live birth rates and insemination method