The concept of recurrent implantation failure (RIF) has been expanded, and the tactics to identify RIF patients have included searching for prognostic criteria of in vitro fertilisation (IVF) outcome, with the aim of minimising the number of unsuccessful attempts and the risk of patient drop-out. However, RIF studies are limited because most suggested methods require endometrial biopsy and subsequently do not allow fresh embryo transfer. Therefore, there is still need for a noninvasive RIF model for prediction of IVF cycle outcome.
Granulocyte colony-stimulating factor (G-CSF) belongs to the family of haemopoietic growth factors. With regard to reproductive physiology, G-CSF has several valuable functions. In women with a normal menstrual cycle, G-CSF leads to leukocyte accumulation in the follicle and accelerates ovulation.1 Given this fact, the administration of recombinant G-CSF was successfully implicated as a preventive tool for luteinised unruptured follicle syndrome.2 Moreover, G-CSF is known to play a mediatory role in the process of oocyte maturation, demonstrating the direct correlation between the number of competent oocytes and the G-CSF level in follicular fluid.3,4 Early studies demonstrated that G-CSF promoted the proliferation of trophoblast cells, thus programming appropriate functioning of the fetal–maternal interface.5 Moreover, G-CSF seems to facilitate embryo competence. Adding G-CSF to embryo culture has benefits, including increasing development and post-transfer survival, as well as decreasing pregnancy loss.6
The aim of the present study was to evaluate whether G-CSF could be used as a reliable prognostic criterion of clinical pregnancy in ‘fresh’ cycles. An open-label, randomised, controlled pilot study with parallel assignment was performed. After obtaining board approval, 83 women <39 years old were recruited. Using blocked randomisation with randomly selected block sizes, the patients were divided into either the study group (n=43) or the control group (n=40).
Matching criteria were RIF, identified as ≥2 unsuccessful IVF attempts; good quality of previously transferred embryos according to the Gardner blastocyst classification; normal karyotype; and the absence of uterine factors of infertility. The age, BMI, and number of unsuccessful IVF attempts did not differ statistically between the groups. The mean number of unsuccessful IVF attempts was four. In the study group on the day of oocyte retrieval, the uterine flushing was collected using an insemination catheter, then marked and frozen. G-CSF concentration in uterine flushing was determined using ELISA and subsequently measurement per gram of protein.
The primary outcome of clinical pregnancy rate was analysed; there was no significant difference comparing between groups (chi-square: 0.018; p>0.05; continuity correction: 0.015). Thus, the method of collection, in this case uterine flushing, did not affect IVF outcome and can be used routinely. The G-CSF concentration in uterine flushing was significantly higher in women with clinical pregnancy with cut-off value of G-CSF being 0.151. The receiver operating characteristic curve showed a sensitivity of 87.5% and a specificity of 94.3%.
The method of collection of uterine flushing does not affect IVF cycle outcome and can be used routinely. According to obtained results, the evaluation of G-CSF level in uterine flushing can be considered a perspective, prognostic criterion for IVF cycle outcome with a sensitivity of 87.5% and specificity of 94.3%. However, as the study was a pilot, the number of recruited participants was limited. Thus, further investigation is needed. Nevertheless, these data prove the necessity for further research into the role of G-CSF in the implantation process and the need to consider the lack of G-CSF as a possible cause of recurrent implantation failure.