Objective
To compare live birth rates following ultrasound-guided embryo transfer (ET) by reproductive endocrinology and infertility fellows versus attending physicians.
Study Design
Women who underwent their first day-3, fresh, nondonor ET between Oct. 1, 2005, and April 1, 2011, at our academic center were included in this retrospective cohort study. Embryos were designated high quality if they had 8 cells, less than 10% fragmentation, and no asymmetry. ET was performed with the afterload technique under ultrasound guidance. Categorical variables were evaluated with the χ 2 test and continuous variables with the Student t test. Logistic regression was performed to assess the relationship between ET physician and live birth rate while adjusting for potential confounders.
Results
Seven hundred sixty women underwent ET by an attending physician, and 104 by a fellow. Baseline characteristics were similar between the groups. The live birth rate was 31% following ET by an attending physician, compared with 34% following ET by a fellow ( P = .65). Logistic regression adjusting for potential confounders demonstrated no significant association between ET physician and live birth rate.
Conclusion
This retrospective study demonstrated no significant difference in live birth rates following ultrasound-guided ET by fellows vs attending physicians at our institution. These data suggest that academic practices using the afterload method and ultrasound guidance can train fellows to perform ET without compromising success rates.
Embryo transfer (ET) is a critical step in the process of in vitro fertilization (IVF). Existing literature suggests that ET outcomes are influenced by several factors, including patient age, embryo quality, the type of transfer catheter used, the use of ultrasound guidance, and ET provider. Despite the importance of proper ET technique, a recent survey of current fellows and recent fellowship graduates indicated that almost half of reproductive endocrinology and infertility (REI) fellows do not perform ET while in training. Although that study was limited by a low response rate of 39%, the findings suggest that many REI fellowship graduates perform their first ETs as attending physicians. Several explanations have been suggested for the lack of experience in ET. Many programs restrict ET to attending physicians based on historical precedent or concerns about patient satisfaction. Alternatively, programs may fear that pregnancy rates will be compromised by allowing fellows to perform ET, despite a lack of supporting evidence. Finally, significant heterogeneity exists among the various training methods. As an alternative to real ET, many programs train fellows with mock ET or intrauterine insemination (IUI). Other programs require fellows to perform a minimal number of IUIs prior to performing ET. In a recent study, however, fellow and attending physician ET pregnancy rates were comparable both before and after the institution of a minimal IUI policy. In addition, a survey of recent fellowship graduates revealed that those who did not perform ET as a fellow were more likely to require additional ET training after graduation. Therefore, there is a critical need for development of a training method that will allow fellows to perform ET while not compromising programs’ success rates.
At our academic institution, fellows perform embryo transfer under direct ultrasound guidance by the attending physician. All providers use the afterload method, in which the outer sheath of the transfer catheter is left in place to maintain access to the uterine cavity. Embryos are loaded into the inner sheath only after proper placement of the outer sheath is confirmed by the attending physician, thereby minimizing the time from loading to transfer and ensuring optimal catheter positioning. Given the high degree of attending control over the ET process at our institution, we hypothesized that our training model allows fellows to perform ET without compromising success rates. Our objective was to examine the relationship between trainee status and live birth rate after ET while controlling for potential confounders.
Materials and Methods
The study was approved by the Northwestern University Institutional Review Board. All women who underwent day 3, fresh, nondonor embryo transfer between Oct. 1, 2005, and April 1, 2011, at our academic center were identified. This study period was chosen based on the inception of our institution’s fellowship program in 2005. Only first IVF cycles were included in the analysis. Two hundred twenty-one cycles lacking complete data on embryo quality, ET physician, or birth outcomes were excluded.
Patients underwent one of the following ovarian stimulation protocols: luteal phase leuprolide acetate suppression (Lupron; Abbott Laboratories, Abbott Park, IL) with or without oral contraceptive pretreatment; microdose Lupron flare; or GnRH antagonist prevention of premature ovulation with cetrorelix (Cetrotide; EMD Serono, Rockland, MA) or ganirelix (Antagon; Organon, Roseland, NJ). Controlled ovarian hyperstimulation was achieved by administration of once or twice daily injections of follicle-stimulating hormone (FSH) (Follistim; Organon, Roseland, NJ or Gonal-F; EMD Serono) or FSH plus luteinizing hormone (LH) (Menopur; Ferring Pharmaceuticals, Parsippany, NJ) at total daily doses ranging from 75 through 600 IU. In the antagonist protocol, the GnRH antagonist was added when a lead follicle measured ≥14 mm or the estradiol exceeded 500 pg/mL. Cycles were monitored with serum estradiol levels and transvaginal ultrasounds beginning on stimulation day 5 and every 1-2 days thereafter. When at least 3 follicles reached a mean diameter of 16 mm, 250 μg recombinant human chorionic gonadotropin (hCG) (Ovidrel; EMD Serono) were administered subcutaneously. Ultrasound-guided oocyte retrieval was performed 36 hours later. Luteal phase support with 50 mg intramuscular progesterone in oil was initiated the day of oocyte retrieval.
Embryos were inseminated 4-6 hours after retrieval by culture with motile sperm or by intracytoplasmic sperm injection. Fertilization was verified by the presence of 2 pronuclei 15 to 18 hours after insemination. Embryos were cultured in 40 μL droplets of culture medium at 37° C in a humidified 5% O 2 , 5% CO 2 , and 90% N atmosphere until day 3. Embryos were deemed high quality on day 3 if they contained 8 cells with less than 10% fragmentation and no asymmetry.
Embryo transfer was performed on day 3. Five fellows and 4 attending physicians performed ET. There was no overlap between fellows and attendings. In general, fellows performed 1-5 intrauterine inseminations and observed 1-5 embryo transfers before performing a transfer; however, there was no minimum requirement for inseminations or observed transfers. The 4 attending physicians each had 15-20 years experience with ET and historically had comparable success rates. Fellows typically performed transfers on 1-2 assigned days per week. The number of embryos to transfer was based on American Society for Reproductive Medicine guidelines and institutional protocols. All ETs were performed with a Wallace catheter (Smiths Medical, Dublin, OH) under direct transabdominal ultrasound guidance. Attending physicians performed the ultrasound guidance for all fellow ETs. The afterload technique was used as previously described by Neithardt et al. Briefly, the cervix was washed with embryo culture media and a Wallace catheter was then introduced into the uterine cavity until the outer sheath passed the internal os. The inner catheter was removed and discarded, and the embryo(s) loaded into a new inner catheter by the embryologist. This inner catheter was then placed through the outer catheter and the embryo(s) was/were transferred approximately 1 to 1.5 cm from the top of the uterine cavity.
Pregnancy was confirmed with a positive serum hCG 10 days following ET, and a repeat hCG 48 hours later if the initial hCG was positive. Clinical pregnancy was verified by fetal cardiac activity on a transvaginal ultrasound at 6 to 7 weeks’ gestational age. Clinical pregnancy rate, live birth rate, and multiple birth rate were calculated as follows: (N/total number of ETs) × 100.
Because of the retrospective study design, sample size was determined by the study period. Previous studies have reported that pregnancy rates with different ET providers vary significantly. For example, Hearns-Stokes et al demonstrated pregnancy rates ranging from 17.0% to 54.3% and Karande et al reported a range of 13.2% to 37.4%. For the current analysis, post hoc power calculations were performed with SPSS Sample Power 3 (IBM Corporation, Armonk, NY). The overall live birth rate for attending physicians and fellows during our study period was 31.7%. For the purpose of power calculations, we hypothesized that the live birth rate would be higher for attending physicians. Assuming a live birth rate of 32% for attending physicians, a sample size of 104 participants per group would provide 83% power at 5% type I error to detect an absolute difference of 17% in live birth rates (ie, 32% for attending physicians and 15% for fellows). Assuming an even higher live birth rate for attending physicians, 35%, the same sample size would still provide 80% power at 5% type I error to detect the same absolute difference (ie, 35% for attending physicians and 18% for fellows).
Statistical analysis was performed with SPSS Statistics 19 (IBM Corporation). The χ 2 test was used for categorical variables, and Student t test for continuous variables. Crude odds ratios (ORs) and 95% confidence intervals (95% CIs) were determined. Logistic regression was performed to examine the association of trainee status with live birth rate while controlling for the effects of potential confounders, including maternal age, gravidity, parity, day 3 FSH, number of oocytes retrieved, number of oocytes fertilized, use of intracytoplasmic sperm injection, use of assisted hatching, number of embryos transferred, and the number of high quality embryos transferred. Only significant predictors of live birth were included in the final regression model. A 2-tailed P value of < .05 was considered statistically significant.
Results
Seven hundred sixty embryo transfers were performed by attending physicians and 104 by fellows. Baseline patient characteristics were similar between the 2 groups ( Table 1 ). There were no significant differences in stimulation parameters, laboratory techniques used, or the number of embryos transferred ( Table 2 ). Embryo quality was similar between the 2 groups ( Table 2 ). There were no significant associations between trainee status and IVF outcomes. Live birth rates, clinical pregnancy rates, and multiple birth rates were comparable between the 2 groups ( Table 3 ). Only young maternal age and the number of high quality embryos transferred were predictors of live birth when all potential confounders were included in a logistic regression model ( P < .001). After adjusting for the effects of maternal age and number of high quality embryos transferred, the odds of live birth remained statistically indistinguishable following ET by an attending physician vs ET by a fellow ( Table 3 ).
| Characteristic | Attending physicians n = 760 | Fellows n = 104 | P value |
|---|---|---|---|
| Age, y | 36.4 (36.1–36.7) | 36.0 (35.2–36.7) | .32 |
| Gravidity | 0.9 (0.8–0.9) | 1.0 (0.7–1.3) | .27 |
| Parity | 0.2 (0.2–0.3) | 0.3 (0.1–0.4) | .70 |
| D3 FSH (mIU/mL) | 7.5 (7.3–7.8) | 7.5 (6.9–8.1) | .91 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
