Characteristics of the study cohort

In this study, a total of 1,009 patients with endometriosis, who were diagnosed and staged via laparoscopy, and underwent their first programmed single FET cycles between January 2018 and April 2024 were included. Among them, 184 cases were excluded due to uterine pathology, thinner endometrium (endometrial thickness of < 7 mm on the day of progesterone administration), using other luteal support protocols and with endocrine diseases and autoimmune diseases. Details were shown in Fig. 1. Finally, 825 FET cycles were included, and 382 pregnancies were achieved, leading to a clinical pregnancy rate of 46.30%. In all programmed FET cycles, 43.88% (362/825) of the patients received vaginal progesterone administration, while the remaining patients (463/825) received intramuscular progesterone administration.

Comparison of baseline characteristics between the two groups

The baseline characteristics were shown in Table 1. Among these patients, the vaginal progesterone group had a shorter infertility duration (2.00 (1.00–4.00) versus 3.00 (2.00–5.00), P < 0.001), higher basal serum LH levels (3.65 ± 1.64 versus 3.37 ± 1.51, P = 0.012), and lower E2 (118.00 (88.00–164.00) versus 130.00 (88.00–178.50), P = 0.002) and progesterone levels (4.20 (2.40–7.50) versus 11.50 (8.60–14.50), P < 0.001) on the day of transfer. However, the two groups were comparable with regards to female age, BMI, gravidity, parity, basal serum FSH and E2, endometriosis r-ASRM stages, blastocyst grade, blastocyst developmental stage, endometrial thickness on the day of progesterone administration, and endometrium preparation protocols (all P > 0.050). In addition, there were no significant differences in clinical pregnancy rates, miscarriage rates, biochemical pregnancy rates, ectopic pregnancy rates, ongoing pregnancy rates, live birth rates and neonatal perinatal outcomes between the two groups (all P > 0.050) (Table 2).

Table 1 Baseline characteristics of patients using vaginal progesterone versus intramuscular progesterone for luteal supportTable 2 Reproductive outcomes of two different routes of progesterone administration

Supplemental Table 1 shows the results of stratification into four groups based on serum progesterone levels on the day of transfer. The results indicate that there were no significant differences in pregnancy outcomes among the different progesterone level groups, both in the whole cohort and when stratified according to the routes of progesterone administration.

Association of different routes of progesterone administration and pregnancy outcomes

A univariate regression analysis was conducted to assess the impact of each variable on clinical pregnancy rate (Supplemental Table 2). Overall, the infertility duration, basal serum FSH levels, and blastocyst developmental stage were negatively correlated with clinical pregnancy rates, while PGT utilization and blastocyst grade were positively correlated with clinical pregnancy rates. Endometriosis r-ASRM stages and the two routes of progesterone administration had no significant impact on clinical pregnancy rates.

Multivariate logistic regression models were used to evaluate the relationship between the two routes of progesterone administration and pregnancy outcomes (Table 3). Clinical pregnancy rates were 49.17% (178/362) in vaginal progesterone group versus 44.06% (204/463) in intramuscular progesterone group, respectively. Miscarriage rate in the vaginal progesterone group was 16.85% (30/178), compared to 24.51% (50/204) in the intramuscular progesterone group. In models unadjusted, partially adjusted, or the fully adjusted, we all found that the differences were not statistically significant in clinical pregnancy rates (aOR 0.82, 95% CI 0.61–1.11, P = 0.197) and miscarriage rates (aOR 1.57, 95% CI 0.90–2.75, P = 0.112) (fully adjusted model) between the two groups.

Table 3 Relationship between different routes of progesterone administration and clinical pregnancy rates and miscarriage rates in different modelsSubgroup analyses

To assess whether the relationship between the two routes of progesterone administration with pregnancy outcomes is stable in different subgroups, we conducted subgroup analyses and interaction tests. The results of the subgroup analysis based on the r-ASRM stages of endometriosis showed that in patients with stages I-II, clinical pregnancy rate was 43.43% in the intramuscular progesterone subgroup, significantly lower than 53.69% in the vaginal progesterone subgroup (aOR 0.74, 95% CI 0.58–0.93, P = 0.011). However, in patients with stages III-IV, there was no significant difference in clinical pregnancy rates between the two groups (aOR 1.02, 95% CI 0.84–1.23, P = 0.855). Additionally, the interaction testing of two routes of progesterone administration and r-ASRM stages was highly significant (P for interaction = 0.036) (Fig. 2). As miscarriage rate was considered, there was no statistically significant difference between the two groups in patients with stages I-II (aOR 1.48, 95% CI 0.99–2.23, P = 0.065) and stages III-IV (aOR 1.11, 95% CI 0.78–1.60, P = 0.559) (Supplemental Fig. 1).

Fig. 2

Effect of different routes of progesterone administration on clinical pregnancy rates in each subgroup. Note: BMI = body mass index, P = progesterone, HRT = hormone replacement treatment, GnRHa = gonadotropin-releasing hormone agonist, PGT = preimplantation genetic testing, r-ASRM = The revised American Society of Reproductive Medicine, OR = odds ratio, CI = confidence interval. aAdjusted for woman’s age at oocyte retrieval, BMI, PGT utilization, blastocyst developmental stage, morphology score, endometrial thickness on the day of P administration, endometrial preparation protocols

We compared the subgroups divided by the r-ASRM stages of endometriosis (Supplemental Table 3). The results showed that subgroup with r-ASRM stages III-IV had a shorter infertility duration (2.00 (1.00–4.00) versus 3.00 (2.00–5.00), P < 0.001), lower ovarian reserve indicated by lower AMH level (2.93 (1.65–4.46) versus 3.56 (2.25–5.26), P < 0.001), higher basal FSH level (6.16 ± 2.04 versus 5.75 ± 1.77, P = 0.003), and lower number of oocytes collected (13.91 ± 7.03 versus 17.41 ± 7.90, P < 0.001). In addition, significant higher proportion of patients in this subgroup received GnRHa downregulation (62.55% versus 36.02%, P < 0.001), resulted in lower LH (0.47 (0.22–3.36) versus 2.96 (0.41–6.36), P = 0.001) and E2 (117.00 (82.75–165.00) versus 131.00 (99.75–188.25), P < 0.001) levels at the embryo transfer day.

We further presented the pregnancy outcomes based on whether GnRHa pretreatment was used, grouped by different r-ASRM stages of endometriosis patients, and further sub-grouped by the two methods of progesterone administration (Supplementary Table 4). The results showed that among these 347 mild patients of endometriosis with stage I-II (222 cases of stage I and 125 cases of stage II), HRT group had a higher clinical pregnancy rate (52.25% vs 40.00%, P = 0.028) and live birth rate (34.23% vs 24.00%, P = 0.047) compared with the GnRHa pretreatment group, and this trend was consistent across both subgroups of progesterone administration methods. However, in severe patients with stage III-IV, the clinical pregnancy rate of the HRT subgroup with vaginal progesterone was significantly lower than that of the GnRHa pretreatment subgroup (37.21% vs 51.97%, P = 0.034), while there was no significant difference in the intramuscular progesterone subgroups and the entire population.

Sensitivity analyses

After addressing the missing data with multiple imputation, the results of the multivariate logistic regression analyses using multiple adjustment strategies were consistent with those of participants who had complete data (Supplemental Table 5).