Effect of growth hormone cotreatment in sub-fertile women ≥ 40 years: A Meta-analysis

ME was responsible for the conception, defining the research question, and study design. ME, LS, and HZ conducted the data, collection, analysis, and interpretation. ME drafted the manuscript. SL and HZ critically reviewed the manuscript. All authors approved the final version of the manuscript.

This meta-analysis did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

One investigator conducted a systematic literature search of the PubMed, Google Scholar, and Cochrane library databases. The medical subject headings of growth hormone cotreatment, growth hormone supplementation, GH, in vitro fertilization, IVF, advanced maternal age, women ≥ 40 years old, clinical pregnancy, live birth, and ovarian stimulation were searched. The investigator used ‘AND’ to filter the search for studies that have combined growth hormone cotreatment or supplementation with ovarian stimulation and/or in vitro fertilization. The search included randomized controlled trials (RCTs) and comparative retrospective trials (CRTs). The investigators were interested in the clinical pregnancy rate per embryo transfer, the number of retrieved and mature oocytes, and the live birth rate. The references of retrieved studies were manually screened to find more candidate studies. The search was initially conducted up to the end of the year 2020; however, a recently published study ( Chen et al. 2022 ) was added. Only studies published in English were included. The three investigators then read the abstracts and methodologies of the filtered studies to identify those that met the inclusion criteria. The investigators emailed the authors of two research studies ( Choe et al. 2018 , Dakhly et al. 2018 ) and are grateful to the authors of one study for making details of the subgroup of women of AMA available ( Dakhly et al. 2018 ). Comparative studies of GH cotreatment vs placebo or no GH. Studies with participants or with a subgroup of women ≥ 40 years old. Studies report clinical pregnancy in women with or without live birth, the number of retrieved oocytes, or their maturity. Co-treatment with GH as the main and only intervention, Couples who underwent non-donor IVF or ICSI treatment with any ovarian stimulation protocol. Investigators excluded cross-sectional and non-comparative studies. The protocol of this meta-analysis with its inclusion criteria, search, and analysis methods was defined in advance and registered in the PROSPERO database ((CRD42021252618). Studies that met the inclusion criteria based on the opinions of two investigators were studied in detail. Three investigators independently reviewed, extracted data, and evaluated the quality of selected studies using a predefined data sheet. The disagreement between the three investigators was resolved by discussion and was only agreed upon with the approval of two investigators. Review Manager 5.4 software (Cochrane Collaboration) was used to analyze the data and formulate the meta-analysis. Dichotomous results were analyzed by calculating the odds ratio (OR) with a 95% confidence interval (CI). The investigators had intended to use the risk ratio but later changed it to the OR after a final script review by a statistician. Continuous variables were studied by calculating the mean difference and 95% Cl. Due to the different GH regimens and protocols used, different ovarian stimulation protocols, fertilization methods, and the day and number of embryos transferred, the investigators preferred to use the random model effect even in the presence of a low degree of heterogeneity, estimated using the I 2 statistic. The publication bias was visually estimated by the funnel plot. The risk of bias was evaluated in the RCTs using the Cochrane Collaboration tool. The Robins-1 assessment tool for non-RCTs was used for CRTs. The absolute difference and the number needed to treat were calculated manually. A P -value less than 0.05 indicates statistically significant results.

The search yielded 406 studies. After reading the abstracts and the methodology of the studies, 399 studies were excluded, as detailed in Fig. 1 . Figure 1 Flow diagram of studies’ selection and inclusion. Flow diagram of studies’ selection and inclusion. Seven studies met the inclusion criteria. Table 1 shows the characteristics of the included studies, while Supplementary Table 1 (see section on supplementary materials given at the end of this article) includes the demography of the studies. Table 2 shows the quality assessment of the CRTs. RCTs’ quality assessments are included in Figures 2 and 3 . Supplementary Table 2 shows details of the bias assessment in the included studies. The investigators were unable to exclude the possibility of publication bias or imprecision due to the asymmetry in the funnel plot (Supplementary Fig. 1A). However, when the study by Ho et al. (2017) ( Ho et al. 2017 ), which included 134 participants, was excluded, the degree of heterogeneity decreased with visual improvement in funnel plot symmetry (Supplementary Fig. 1B) without affecting the significance of the measured improvement in clinical pregnancy rate in the co-treatment group with GH (OR: 2.77 (95% CI): 1.75–4.39; I 2 = 0%). Figure 2 (A) Clinical pregnancy per fresh embryo transfer in women ≥ 40 years old who used/did not use GH cotreatment in their IVF/ICSI trial. (B) Absolute risk difference of the clinical pregnancy per fresh embryo transfer in the subgroup of women ≥ 40 years old who used/did not use GH cotreatment in their IVF/ICSI trial and were classified as patients with poor response and good response. Figure 3 (A) Live birth per fresh embryo transfer in women ≥ 40 years old who used/did not use GH cotreatment in their IVF/ICSI trial. (B) Risk difference of the live birth per fresh embryo transfer in women ≥ 40 years old who used/did not use growth hormone cotreatment in their IVF/ICSI trial. Table 1 Characteristics of the included studies. Reference Study period Study design Participants and inclusion and exclusion criteria Intervention ovarian stimulation and GH dose and duration Outcome Tesarik et al. (2005) - RCT 100 women aged 41 to 44 years. Couples with azoospermia and women with basal FSH > 14 IU/mL or those with basal inhibin B of <30 pg/mL (<110.1 pmol/L) were excluded. Daily s.c. injection of 8 IU of GH started on day 7 of stimulation until the day of ovulation trigger. The control group received only the solvent. All were non-donor cycles. Primary outcome: delivery and live birth rates Collected for this meta-analysis. Clinical pregnancy rate. Live birth rate. Number of retrieved oocytes. Number of MII oocytes. Keane et al. (2017) 1 April 2008 to 31 December 2015 Retrospective 163 women aged ≥40 years who had failed at least 1 IVF (a subgroup of the total 400 women). Only the first cycle was considered in the analysis. Excluded were freezing all cycles, canceled cycles, failed fertilization, failed oocyte retrieval, pregnancy with ectopic and blight ovum, and donor cycles. Some women received GH (Saizen) injections starting in the previous menstrual cycle (from day 2 or 3) and included 6 injections over 6 weeks leading up to the day of oocyte retrieval. Approximately 54 IU were administered for 33–37 days (1.5 IU per day). Others received 1 IU of GH injection (Sci-Tropin) per day for 45 days up to the day of oocyte retrieval. Primary outcome: clinical pregnancy and live birth rate Collected for this meta-analysis. Clinical pregnancy rate. Live birth rate. Ho et al. (2017) January 2005 to December 2009 Retrospective 134 women aged ≥40 years old. No exclusion criteria were mentioned. 3 IU daily injections of GH from day 3 coincide with the start of the gonadotrophin injection and up to the day of triggering. Fresh embryo transfer with at least two blastocysts of the best quality. Primary outcome: number of oocytes and embryos, quality of embryos, and implantation and pregnancy rates Collected for this meta-analysis. Clinical pregnancy rate. Number of retrieved oocytes. Number of MII oocytes. Lan et al. (2019) January 2009 to March 2014 Retrospective Women ≥ 40 years old with a history of poor ovarian response or poor ovarian reserve (POSEIDON group 4) who underwent their first IVF cycles in their center were included. Women with a history of intrauterine synechiae, congenital Mullerian duct anomaly, hydrosalpinx, endometrial fluid, sub-mucosal myoma, and incidentally found endometrial lesions during ovarian stimulation were excluded. 15 women (9 with and 6 without GH supplementation), were excluded. 8 IU of daily injections of GH starting on the day that the leading follicle had reached 14 mm in diameter until the day of triggering. The authors evaluated the reproductive outcome of IVF/ICSI. Collected for this meta-analysis. Clinical pregnancy rate. Number of retrieved oocytes. Lee et al. (2019) January 2010 to October 2012. Combined RCT and retrospective The study was divided into two parts. The first part was RCT and was included in this meta-analysis. The inclusion criteria included women classified as poor responders according to the definition of the Bologna criteria. Only the first IVF cycle with the fresh transfer was managed by the same clinician. A total of 10 IU of GH were administered, divided into 4, 4, and 2 IU of GH daily injections for 3 consecutive days along with induction of ovulation. Primary outcome: implantation rate, clinical pregnancy rate, miscarriage rate, and ongoing pregnancy rate. Collected for this meta-analysis. Clinical pregnancy rate. Chen et al. (2022) June 2018 to December 2019 Retrospective Couples with unexplained poor embryonic development after previous IVF (i.e. no top-quality embryos graded 1 or 2). Inclusion criteria were BMI between 18 and 25 kg/m 2 , tubal infertility, regular menstrual cycle, normal uterine morphology, and no residual frozen embryo. The exclusion criteria were preimplantation genetic testing cycle (PGT), endometriosis, polycystic ovary syndrome, medical comorbidities, azoospermia, and the insemination method changed in the second IVF cycle. In the antagonist group (approximately 75%) of the patients, 3 IU daily injections of GH (Jintropin AQ) were started from the day of the start of the gonadotropin injection until the day of the hCG (about 10 days). In the agonist group (about 25) of the patients, 2 IU of daily injections of GH (Jintropin AQ) were started after the pituitary down-regulation was confirmed and until the day of hCG injection (about 25 days) Primary outcome: the oocyte–cleavage rate and the clinical pregnancy rate Collected for this meta-analysis. Embryos cleave rate. Clinical pregnancy rate. Dakhly et al. (2018) April 2015 to November 2017 Open-label RCT Women who met the Bologna criteria. The authors excluded women > 45 years old, basal FSH > 20 IU/mL, or husbands with azoospermia or severe teratozoospermia. GH cotreatment 2.5 mg s.c. injection (equivalent to 7.5 IU) (Norditropin pen, Novo Nordisk, Denmark) from day 21 of the previous cycle to the day of hCG. Primary outcome: the live birth rate (fresh, frozen, and cumulative). Collected for this meta-analysis. Clinical pregnancy. Live birth. Number of oocytes. Number of mature oocytes. Table 2 Quality evaluation of retrospective studies. Robins-1 tool for evaluating nonrandomized controlled trials. Study Preintervention At intervention Postintervention Overall risk Confounding bias Selection bias Intervention bias Deviation from the intended intervention Missing data Outcome management Selective reporting bias Ho et al. (2017) Serious Low Low ? ? Low ? Serious Keane et al. (2017) Moderate Moderate Low Low Low Low Low Moderate Lan et al. (2019) Moderate Moderate Low Low Low Low Low Moderate Chen et al. (2022) Serious Low Moderate Low Serious Low Low Serious ?: could not be assessed due to lack of data (A) Clinical pregnancy per fresh embryo transfer in women ≥ 40 years old who used/did not use GH cotreatment in their IVF/ICSI trial. (B) Absolute risk difference of the clinical pregnancy per fresh embryo transfer in the subgroup of women ≥ 40 years old who used/did not use GH cotreatment in their IVF/ICSI trial and were classified as patients with poor response and good response. (A) Live birth per fresh embryo transfer in women ≥ 40 years old who used/did not use GH cotreatment in their IVF/ICSI trial. (B) Risk difference of the live birth per fresh embryo transfer in women ≥ 40 years old who used/did not use growth hormone cotreatment in their IVF/ICSI trial. Characteristics of the included studies. Clinical pregnancy rate. Live birth rate. Number of retrieved oocytes. Number of MII oocytes. Clinical pregnancy rate. Live birth rate. Clinical pregnancy rate. Number of retrieved oocytes. Number of MII oocytes. Clinical pregnancy rate. Number of retrieved oocytes. Clinical pregnancy rate. Embryos cleave rate. Clinical pregnancy rate. Collected for this meta-analysis. Clinical pregnancy. Live birth. Number of oocytes. Number of mature oocytes. Quality evaluation of retrospective studies. Robins-1 tool for evaluating nonrandomized controlled trials. ?: could not be assessed due to lack of data All included studies investigated GH as the main and only intervention. The total number of participants included was 881, including those from 2 studies that included only women ≥ 40 years old ( n  = 357) and those from 5 studies that included sub-group of women ≥ 40 years old ( n  = 424) where the data of this sub-group were extracted. All studies compared GH cotreatment in IVF/ICSI with fresh embryo transfer cycles with GH (+) or without GH/placebo (−). A pooled analysis of included studies of AMA women who underwent IVF/ICSI with fresh embryo transfer showed a significant increase in both clinical pregnancy and live birth rates without a significant difference in the number and maturity of the retrieved oocytes. The clinical pregnancy rate per fresh embryo transfer was reported in 7 studies that included 881 patients (481 received GH treatment and 400 without). However, one of these studies ( Dakhly et al. 2018 ) had double zero events and was excluded from the OR calculation but was included in the calculation of the pooled absolute risk difference (ARD) and the number needed to treat (NNT). There was a statistically significant increase in clinical pregnancy per embryo transfer (OR: 2.2; (95% CI): 1.34–3.61; I 2  = 31%) ( Fig. 2 ). After excluding four non-RCTs ( Ho et al. 2017 , Keane et al. 2017 , Lan et al. 2019 , Chen et al. 2022 ), the clinical pregnancy rate was still significantly higher in GH (+) with (OR: 4.48; (95% CI): 1.58–12.64; I 2  = 0%) while excluding studies with a high risk of bias ( Ho et al. 2017 , Lee et al. 2019 , Chen et al. 2022 ) OR of 2.27; (95% Cl: 1.52–4.66; I 2  = 0%) was still statistically significant. The calculated weighted risk difference (RD) of the seven studies (GH (+) 107/481 and GH (−) 44/400) was 0.09; 95% CI: 0.03–0.14; I 2  = 37% ( Fig. 2B ). Based on weighted RD, if the chance of clinical pregnancy is 11% (44/400) in GH (−), it would be between 14% (11% + 3%) and 25% (11% + 14%) with GH (+). The NNT is 8.9 (1/(107/481) − (44/400)). A subgroup analysis was performed in AMA women with poor response to ovarian stimulation as defined by Bologna criteria ( Ferraretti et al. 2011 ). The results of the 3 studies ( Keane et al. 2017 , Dakhly et al. 2018 , Lee et al. 2019 ) with 273 participants did not show significant ARD 0.05 (95% CI: −0.02 to 0.12; I 2  = 25%) ( Fig. 2B ). Other studies were considered to include good responder patients (608 participants). Pooled data analysis showed a statistically significant ARD of 0.12 (95% CI: 0.04–0.20; I 2  = 34%). Based on weighted RD, if the chance of clinical pregnancy in the good responder group is 13.7% (37/270) in GH (−), it would be between 17.7% (13.7% + 4%) and 33.7% (13.7% + 20%) with GH (+). The NNT is 7.9 (1/(89/338) − (37/270)). ( Fig. 2B ). Good responders had an average retrieved oocyte of ≥4.2 ± 2.3. Four studies ( Tesarik et al. 2005 , Keane et al. 2017 , Dakhly et al. 2018 , Chen et al. 2022 ) reported the rate of live birth per fresh embryo transfer. They included 431 women, 204 in the GH (+) group and 237 in the control group GH (−). One study ( Dakhly et al. 2018 ) had double zero events and was not included in the analysis. A pooled analysis of three studies ( Fig. 3 ) showed a significantly higher live birth rate (OR: 4.12; 95% CI: 1.82–9.32; I 2  = 0%). The pooled calculated ARD of four studies (GH (+) 28/196 and GH (−) 9/223) was 0.08; 95% CI: 0.01–0.16; I 2  = 50% ( Fig. 3B ). Based on that, if the chance of clinical pregnancy is 4% (9/223) with GH (−), it would be between 5% and 20% with GH (+). The NNT is 9.75 (1/((28/196) − (9/223)). This meta-analysis did not show a statistically significant difference in the number of mature and retrieved oocytes between GH (+) and GH (−). The OR of the average number of retrieved oocytes was 0.14; 95% CI: −0.75 to 1.03; I 2 = 81%. The total of participants in the 5 included studies ( Tesarik et al. 2005 , Ho et al. 2017 , Dakhly et al. 2018 , Lan et al. 2019 , Chen et al. 2022 ) was 749. The total number of participants in the 4 studies ( Tesarik et al. 2005 , Ho et al. 2017 , Dakhly et al. 2018 , Chen et al. 2022 ) that reported an average number of mature oocytes was 492. The OR was 0.87; 95% CI: −1.8 to 0.06; I 2 = 83% (Supplementary Fig. 2).

This meta-analysis showed a significant increase in both clinical pregnancy and live birth rates in women in AMA who used GH cotreatment and underwent IVF/ICSI with fresh embryo transfer. However, subgroup analysis showed that this occurs only with good responder patients. There was no difference in the number of retrieved oocytes or their maturity. A recently published Cochrane systematic review reported a significant increase in clinical pregnancy but not in a live birth in women aged > 40 years ( Sood et al. 2021 ). Their conclusion was based on four RCTs; however, two of the included studies had recruited women with an average age below 40 years ( Choe et al. 2018 , Dakhly et al. 2018 ). The investigators thank the authors of one of these studies ( Dakhly et al. 2018 ) who confirmed that they did not have a clinical pregnancy in the GH (+) or GH (−) groups in women of AMA (personal communication). Unfortunately, we did not receive a response from the authors of the second paper. It is well-accepted that women of AMA have a lower quantity and quality of oocytes; therefore, poor-quality embryos with a low chance of pregnancy ( Wyns et al. 2021 ). The meta-analysis showed a non-significant difference in the number of retrieved oocytes in women of AMA who received GH. This suggests that GH may improve the competence of oocytes. In fact, GH has been reported to be an important dynamic requirement for folliculogenesis ( Devesa & Caicedo 2019 ), and a higher level of follicular GH was associated with a better quality of oocytes and embryos ( Mendoza et al. 2002 ). GH supplementation increases the mitochondrial DNA copy number of granulosa cells ( Kölle et al. 2004 , Weall et al. 2015 , Li et al. 2020 ), which is correlated with better embryo quality and higher implantation rate ( Ogino et al. 2016 ). However, the mechanisms of the action of GH are complex and are not fully understood. Liu et al. (2020) reviewed the effect of GH on the endometrium and reported that most studies showed positive effects on endometrial receptivity but highlighted the need for more studies ( Liu et al. 2020 ). In addition to the moderate degree of heterogeneity (I 2 of 31%, Chi 2 P  = 0.2) among included studies that evaluated clinical pregnancy per embryo transfer and the low heterogeneity (I 2 of 0%, Chi 2 P  = 0.57) among live birth reporting studies, there were considerable clinical variabilities. The dose and regimen of the GH used were different between studies, with some authors using it during the gonadotrophin stimulation, while others used it during the previous menstrual cycle ( Keane et al. 2017 , Dakhly et al. 2018 ). Even among those who used it during ovarian stimulation, the dose and the start day of GH injection were different. Not all women of AMA are the same, some may have had repeated implantation or pregnancy failure or have never been pregnant with a long duration of infertility. Studies ( Keane et al. 2017 , Dakhly et al. 2018 , Lee et al. 2019 ) that included women of AMA with poor response to ovarian stimulation according to Bologna criteria ( Ferraretti et al. 2011 ) reported no statistically significant differences between the GH (+) or GH (−) groups. The subgroup analysis showed that women of AMA who met the Bologna criteria did not benefit from co-treatment with GH. However, those who responded well to ovarian stimulation showed a significant increase in clinical pregnancy. This should be expected as GH does not increase ovarian reserve. With the advancement of women's age, oocyte quality and quantity decrease. The investigators hypothesize that improvement of the oocyte competency by GH in poor responder women of AMA could not overcome the deterioration of their ovarian reserve. As there is more than a 7% trend of higher clinical pregnancy rates in poor responder women of AMA who used GH, it is not sure whether a different dose or regimen of GH may favorably impact the outcome in this sub-group of women. The number of women fulfilling the criteria of poor responder was small ( n  = 273), and hence, statistical power may not be adequate. Studies of larger sample sizes may refute or confirm this finding. The exception in the good responder group was the study by Ho et al. (2017) . In this study, the average number of retrieved oocytes was more than five. However, the authors did not report significant differences in clinical pregnancy between the GH (+) and GH (−) groups. They argued that it could be due to the use of 3 IU of GH injections per day in their study and that the higher dose (8 IU) used in the study by Tesarik et al. (2005) may explain the benefit in the reported pregnancy outcome. In the study by Ho et al. , only one or two blastocysts were transferred compared to the higher number of cleaved stages transferred embryos in the study by Tesarik et al. Furthermore, the mean basal FSH was 4.6 ± 2.4 in GH (+) and 4.2 ± 2.5 in GH (−). This is compared with participants from Tesarik et al. who had a baseline FSH of 10.2 ± 1.9 in the GH (+) group and 10.1 ± (1.8) in the placebo group. The investigators were unable to find a dose-finding study or suggest a dose or regimen of GH. Furthermore, the possibility of introducing harm with cotreatment with GH could not be excluded. Excessive proliferation of theca and stromal cells due to a high dose of GH had a negative impact on the survival of the prenatal follicle. However, this finding should be taken with caution, not only because it was in a rat ovary but also because it was an in vitro study, and the finding could be attributed to the depletion of the nutrient in the medium ( Zhao et al. 2000 ). The included studies reported that they did not encounter serious side effects or did not mention them. Women of AMA are at increased risk of having diabetes. Co-treatment with GH may not be suitable for diabetic patients due to its negative effect on insulin resistance and glucose blood level ( Jeffcoate 2002 ). Animal studies have shown that GH is oncogenic ( Perry et al. 2006 ) and may be related to a more aggressive form of endometrial cancer ( Pandey et al. 2008 ). However, a systematic review showed that long-term use of GH was not associated with an increased risk of malignancy ( van Bunderen et al. 2014 ). Further studies should address the safety of GH cotreatment in IVF/ICSI cycles. In some studies, patients were allowed to choose whether to use GH or not. Depending on the brand of medication, dose, and regimen used, it is estimated that GH cotreatment can cost between $70 and $460 per treatment cycle. This, in addition to other reasons, can affect the patient's choice and introduce bias. Further studies should consider this. All studies did not mention the cause and duration of infertility. The authors of all studies used a long-down regulation agonist protocol, but in some studies, flare-agonist and antagonist protocols were also used. The dose of daily gonadotropin injections ranged from 150 to 600 IU. Although the included studies recruited women of AMA, women over 45 years of age were hardly included. This meta-analysis suggests that not all women of AMA would benefit from co-treatment, particularly poor responders, according to Bologna criteria. However, this conclusion was based on only 3 studies with 273 patients. Furthermore, the investigators were unable to suggest an ovarian reserve test cut-off level that may justify the use or not of GH cotreatment. Further RCTs are needed. The use of different doses and regimens of GH, different primary outcomes, and the restriction of this meta-analysis to studies published in English are the main limitations. Other limitations are due to the small sample size and significant quality variation between included studies. Confounding factors, selection bias, and missing data were recognized as causes of moderate or serious biases ( Tables 2 and Supplementary Table 2). Furthermore, the assumption that GH cotreatment would improve oocyte competency and/or endometrial receptivity in women of AMA should be taken with caution, as the result of this meta-analysis has the limitation of not being per cycle or with cumulative live birth. The investigators could not find a study with cumulative live birth as an outcome except the study by Dakhly et al. that reported zero live birth in AMA women ( Dakhly et al. 2018 ). The meta-analysis included retrospective studies and an open-label RCT that can exaggerate the effect of the intervention. The asymmetry of the funnel plot suggests publication bias, which, without a change in statistical conclusion, was visually improved by excluding one study. The investigators performed a sensitivity analysis excluding the results of low-quality and nonrandomized studies. Although the result did not show a significant change, the findings of this meta-analysis are undoubtedly far from conclusive. More RCTs are needed, particularly those considering cumulative live birth per start cycle as the main outcome.

This meta-analysis showed a significant improvement in clinical pregnancy and live birth rates with the use of GH cotreatment in a subgroup of women of AMA, but due to the unclear dosage and regimen of GH, it is difficult to confidently recommend its routine use. Further studies should consider the optimal dose and duration of GH injection and its safety.

Animal studies showed that growth hormone (GH) stimulates early follicular growth, improves antrum formation, modifies the growth of developing follicles, stimulates preantral and small antral follicles that lead to the development of healthy granulosa cells, more mature oocytes, and a better fertilization rate ( Magalhães et al. 2011 , Araújo et al. 2014 , Serafim et al. 2015 ). Most GH receptors (GHR) are found in the liver and also in the ovary, including the human ovary ( Abir et al. 2008 ). GHR mRNA expression was found in all stages of follicular development and stromal, mural, and cumulus cells ( Martins et al. 2014 ). GH acts directly through its receptors or indirectly through the insulin-like growth factor system to enhance the action of follicle stimulation hormone (FSH) in granulosa cells ( Silva et al. 2009 ), leading to a higher level of estradiol production ( Araújo et al. 2014 , Serafim et al. 2015 ). In women with diminished ovarian reserve, GH supplementation increased the expression of GH, FSH, and luteinizing hormone (LH) receptors in granulosa cells ( Regan et al. 2018 ). Using GH cotreatment, Regan et al. (2018) randomly and blindly recruited women to receive 60 IU of GH injections in six divided doses. They included those with a history of recurrent implantation failure and poor responders. In the subgroup of women of advanced maternal age (AMA), GH cotreatment improved the density of the FSH, bone morphogenic protein 1B, and LH receptors of the large follicles, which may explain the improvement in the result of live birth documented by the authors ( Regan et al. 2018 ). Weall et al. (2015) reported an increase in mitochondrial activity and an improvement in oocyte quality in women who received GH ( Weall et al. 2015 ). Others confirmed the increased mitochondrial DNA copy number in the oocytes of patients with poor embryonic development and an improvement in the rate of live birth ( Li et al. 2020 ). GH stimulates glucose utilization in embryos and reduces lipid accumulation, which can improve embryo metabolism and quality ( Kölle et al. 2004 ). Furthermore, GH improves the hatching of animal blastocysts ( Drakakis et al. 1995 ). The uterus produces and responds to GH, which promotes uterine growth ( Oliveira et al. 2008 ), enhances endometrial receptivity, and improves endometrial blood flow ( Xue-Mei et al. 2016 , Altmäe et al. 2017 ). GH has been used for many years in ovulation induction ( Homburg & Ostergaard 1995 ), women with polycystic ovarian disease ( Homburg et al. 1995 ), poor responders ( Cozzolino et al. 2020 ), and women with a thin endometrium ( Altmäe et al. 2017 , Cui et al. 2019 ) without uniform agreement on its benefit ( Hart 2019 , Zhu et al. 2020 ). To assess the effect of GH cotreatment in women of AMA, the investigators conducted a systematic review and meta-analysis of GH cotreatment studies in women in AMA who underwent IVF/intra cytoplasmic sperm injection (ICSI) using their autologous oocytes.

The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.