Methods
We enrolled women who experienced miscarriage following fresh or frozen cycles at the Center for Reproductive Medicine, Department of Obstetrics and Gynecology, Peking University Third Hospital between January 2015 and December 2022. The first cycle was included if the patient underwent multiple cycles during the time. Exclusion criteria were listed as following: (1) the gestational week of miscarriage undergoing vacuum aspiration over 16 weeks; (2) experiencing spontaneous miscarriage (spontaneous removal of products of conception); (3) undergoing drug-induced abortion; (4) missing IPI data or unclear miscarriage management.
Our centre had previously documented the in vitro fertilisation (IVF) procedures in detail.11 To summarise, fertilisation was achieved through either traditional IVF or intracytoplasmic sperm injection. Our centre’s specific methods for embryo vitrification and thawing had been outlined in prior research.12 13 We transferred up to two vitrified embryos, either at the cleavage stage or blastocyst stage, during a natural, stimulated or artificial hormonal replacement cycle. In fresh cycles, endometrial thickness was measured via transvaginal ultrasound around the time of the luteinising hormone surge, with a thickness of <7 mm considered as a thin endometrium. A previous study had clearly defined what constituted a good-quality embryo.14 To confirm biochemical pregnancy, we routinely measured serum β-human chorionic gonadotropin approximately 14 and 21 days postembryo transfer. Clinical pregnancy was defined as the presence of an intrauterine gestational sac found by transvaginal ultrasound performed approximately 28–35 days after embryo transfer.
In this study, the IPI was defined as the time from the most recent vacuum aspiration procedure to the subsequent embryo transfer. A miscarriage was defined as the spontaneous end of a clinical pregnancy before the completion of 24 weeks of gestation, where the fetus(es), being non-viable, did not get naturally absorbed or expelled from the uterus. An experienced obstetrician-gynaecologist made the diagnosis of miscarriage by using transvaginal ultrasound, which involved assessing specific ultrasound indicators that suggested the absence of an intrauterine gestational sac or fetal heartbeat. Before diagnosing a miscarriage, it was imperative to confirm the presence of an intrauterine pregnancy through ultrasound. In our study, the gestational age was determined based on the timing of the surgical procedure, as the exact date of the miscarriage diagnosis was not documented. Prior to the surgical procedure, women underwent the necessary preoperative assessments, such as blood tests and vaginal infection screenings. These preparatory steps, while essential, could lead to a delay between the diagnosis of miscarriage and the actual surgical procedure, resulting in a higher gestational week being recorded at the time of surgery.
All patients diagnosed with miscarriage underwent vacuum aspiration with or without hysteroscopy. The details of the procedures have been described previously.12 In summary, all women underwent vacuum aspiration, with controlled cervical dilation as deemed necessary. The decision to undergo hysteroscopy was made primarily based on patient preference after thorough counselling regarding the potential benefits and risks of the procedure. For women undergoing vacuum aspiration only, a rigid vacurette cannula facilitated efficient aspiration. For women undergoing vacuum aspiration plus hysteroscopic surgery, hysteroscopy was conducted to thoroughly remove any residual pathological tissue. A Bettocchi hysteroscope, accompanied by the necessary accessories, ensured clear visualisation and precise manual excision of tissue while preserving as much healthy endometrium as possible. Postprocedure, the uterus was visually inspected to confirm complete emptying. Following the surgery, patients were prescribed hormone replacement therapy and a course of antibiotics. Routine ultrasound examinations were scheduled approximately 1 month later, with surgical re-intervention conducted if any residual pregnancy material was observed in the uterine cavity. Women were then scheduled for a new FET after a minimum of two menstrual cycles had elapsed.
The primary outcome of the study was live birth, defined as the birth of at least one live infant after 24 weeks of gestation.15 The secondary outcomes included several key indicators related to pregnancy and infant outcomes. Biochemical pregnancy was defined as a positive pregnancy test, indicating the presence of pregnancy-related hormones in the body. Clinical pregnancy was defined as the presence of at least one gestational sac on ultrasound at approximately 7 weeks of gestation, with the detection of heartbeat activity. Multiple pregnancy was defined as the presence of at least two gestational sacs on ultrasound at approximately 7 weeks of gestation, with the detection of heartbeat activity. Miscarriage was defined as the spontaneous loss of a clinical pregnancy before 24 completed weeks of gestational age. For women who had an ongoing pregnancy, additional outcomes were reported. These included preterm delivery, defined as an infant born before 37 weeks of gestation. The sex of the newborn was also recorded. Low birth weight was defined as a birth weight of <2500 g at any gestational week, while macrosomia was defined as a birth weight of >4000 g at any gestational week. Small for gestational age and large for gestational age were defined using the reference birth weight percentiles as less than the 10th percentile and more than the 90th percentile, respectively, using two different birth weight references in China.16 17 Caesarean section was also reported as an outcome.
All participants were followed up approximately 2 months after delivery by a trained nurse through a phone call. The details of the follow-up were well described in a previous study.18 This comprehensive approach allowed for the collection of a wide range of data on fertility and neonatal outcomes, which can be used to inform clinical practice and improve patient care.
Statistical analysis
Continuous variables were presented differently based on their distribution: means (SD) for normally distributed data and medians (25th–75th percentiles) for non-normally distributed data. Categorical variables were represented as frequencies (percentages). For comparing continuous variables, Student’s t-test was used for normally distributed data, while the Mann-Whitney U test was employed for non-normally distributed data. The χ2 test was used for comparing categorical variables. Logistic regression models were used to compute ORs with 95% CIs, adjusting for potential confounders, including female age, body mass index (BMI), primary infertility, endometrium thickness, uterine anomalies, recurrent miscarriage history, intrauterine adhesion history, ovulatory dysfunction, endometriosis, fertilisation method, oocyte number, IVF duration, curettage methods and endometrium preparation. Neonatal outcomes were only assessed in women with singleton pregnancies. Additionally, restrictive cubic spline (RCS) regression models were applied to explore the relationship between IPI (treated as a continuous variable) and both fertility and neonatal outcomes. Missing data on male age, primary infertility, infertile duration, BMI, oocyte number and endometrium thickness were observed in 0.2%, 0.2%, 4.4%, 10.7%, 3.4% and 4.9% of the overall cohort, respectively. These missing values were imputed using fully conditional specification compatible with substantive models.19 A sensitivity analysis was conducted to assess whether the results differed in the non-imputed data. Statistical significance was defined as a two-tailed p value <0.05 in all analyses. The analyses were conducted using R software (V.4.3.1).
