Baseline characteristics of patients
No significant differences concerning baseline characteristics were evident between the 2 groups of patients completing the study (Table 1). Among the 32 Hishi patients, 8, 22, 2, and 28 patients respectively had tubal infertility, ovarian dysfunction, endometriosis, and male infertility, as did 5, 21, 1, and 28 of the 31 control patients, showing no significant differences in infertility causes (chi-squared test). Five women in each group were diagnosed with PCOS, representing similar frequency (16% for each group, p = 1.00; Fisher’s exact test). Glucose tolerance was similar between the 2 groups (chi-squared test), including values that were normal, borderline, and diagnostic for diabetes mellitus in 30, 2, and 0 Hishi patients and in 31, 0, and 0 control patients, respectively. Both groups showed normal serum concentrations of hepatic enzymes and thyroid hormones, with no significant differences between the groups.
In summary, patients who completed the study were endocrinologically and metabolically normal infertile women at ages 38 to 42 years, showing no significant differences between groups in ovarian reserve, potential infertility causes, insulin sensitivity and secretion, or glucose metabolism.
Sixty-four ART patients were enrolled and randomized to Hishi and control groups (32 patients each). All Hishi patients completed the study; 1 control dropped out for unknown reasons, leaving 31 (Fig. 1). Both groups underwent conventional infertility treatments in the initial cycle, respectively resulting in 2 live deliveries and 1 miscarriage. Two deliveries in the Hishi group were obtained by sexual intercourse, with spontaneous ovulation in one patient and with ovulation following an LH surge induced by GnRHa after rFSH stimulation in the other. In the control patient, intrauterine fetal death occurred at 8 weeks following ovarian stimulation with clomiphene citrate and intrauterine insemination.
The remaining 30 patients in each group underwent ovarian stimulation, followed in all by successful oocyte retrieval. Subsequently, 27 and 29 fresh ET were carried out in Hishi and control groups respectively, resulting in 6 and 4 pregnancies with 4 and 3 live deliveries. Pregnancy and delivery rate per fresh ET tended to be higher in the Hishi group (22% and 15%) than in controls (14% and 10%) although statistical significance was not attained (p = 0.50 and p = 0.70, Fisher’s exact test; Fig. 2A). Numbers of transferred embryos were similar between Hishi and control groups (2.0 ± 0.3 and 2.0 ± 0.6, p=0.99, Mann-Whitney U test).
Among 17 Hishi and 15 control patients, 20 cycles of cryopreserved ET were carried out in each group, resulting in 9 and 3 pregnancies and 9 and 2 live deliveries, respectively. Hishi extract strikingly increased rates of pregnancies and live deliveries for cryopreserved ET compared with outcomes in controls (45% vs. 15%, p<0.05, relative risk [RR] 4.6, 95% confidence interval [CI] 1.02 – 21.0 and 45% vs. 10%, p<0.05, RR 7.4, 95% CI 1.3 – 40.5; chi-squared test; Fig. 2A). Numbers of transferred embryos did not differ significantly between Hishi and control groups (1.9 ± 0.4 and 1.9 ± 0.4, p=1.00, Mann-Whitney U test). Implantation rate per embryo also was much higher in the Hishi group (35% vs. 8.1%, p<0.01, RR 6.1, 95% CI 1.6 – 23.9; chi-squared test; Fig. 2B).
Next we compared fresh and cryopreserved ET within groups. In Hishi patients, implantation rate per embryo was significantly higher in cryopreserved than fresh ET (35% vs. 15%, p < 0.05, RR 3.0, 95% CI 1.1 – 8.4; chi-squared test; Fig. 2B). In controls, no significant difference was evident between cryopreserved and fresh ET (8.1% vs. 8.8%, p = 1.00; Fisher’s exact test). With Hishi, delivery rate per ET similarly was significantly higher for cryopreserved than fresh ET (45% vs. 15%, p < 0.05, RR 4.7, 95% CI 1.2 – 18.7; chi-squared test, Fig. 2A), compared with no significant difference between ET types in controls (10% vs. 10%, p = 1.00; Fisher’s exact test). These results suggest that Hishi enhances endometrial receptivity and potential to maintain pregnancy, especially in natural cycles rather than controlled ovarian hyperstimulation (COH).
Oocyte potential for fertilization and embryonic development
No significant differences were evident between Hishi and control groups (n = 30 in each) in number of follicles larger than 16 mm in diameter (6.5 ± 3.7 and 6.7 ± 3.6, respectively), serum E2 (3410 ± 2470 pg/mL and 3470 ± 2020), or endometrial thickness (11.2 ± 2.0 mm and 10.8 ± 3.6) on the day of hCG administration (Mann-Whitney U test). Numbers of retrieved oocytes (13.6 ± 10.4 and 15.7 ± 9.4) and mature oocytes (6.1 ± 6.2 and 5.9 ± 4.4) did not differ significantly between Hishi and control groups (Mann-Whitney U test).
Developmental potentials of oocytes and embryos, however, were significantly greater in the Hishi group (Fig. 2C). Oocytes were significantly more likely to develop into fertilized oocytes, day-2 embryos, or superior day-2 embryos in the Hishi group than in controls. The Hishi group also showed a significantly higher rate of fertilized oocytes developing into superior day-2 embryos. Further, significantly more day-2 embryos as well as day-3 embryos showed favorable morphology in the Hishi group.
Embryo transfers were carried out at 2, 3, and 5 days after oocyte retrieval in 2, 12, and 13 Hishi subjects and in 3, 8, and18 control subjects, respectively. The distribution of ET days did not differ significantly between the 2 groups (chi-squared test).
Data from IVF and ICSI were analyzed together since no significant differences were present (Mann-Whitney U test) in the number of fertilized oocytes (9.6 ± 8.4 vs. 7.6 ± 5.9, respectively) or day-2 embryos (8.8 ± 8.0 vs. 6.4 ± 6.0) between IVF and ICSI. Further, delivery rates did not differ significantly among IVF, ICSI, and IVF+ICSI patients (9.8, 9.1, and 25%, respectively; chi-squared test).
Overall study outcomes
In 32 Hishi vs. 31 control patients completing the study, 15 vs. 5 live deliveries were achieved by either conventional infertility treatments (2 vs. 0 deliveries) or ART (13 vs. 5). Thus, cumulative frequency of live delivery per patient was 47% in the Hishi group, significantly higher than 16% in controls (p<0.01, RR 4.6, 95% CI 1.4 – 15.0, chi-squared test; Fig. 2D). In ART including both fresh and cryopreserved ETs, live delivery rate per ET and implantation rate per embryo both were significantly higher in the Hishi group than in controls (28% vs. 10%, p<0.05, RR 3.4, 95% CI 1.1 – 10.4 and 23% vs. 8.5%, p<0.01, RR 3.3, 95% CI 1.4 – 7.8; chi-squared test; Fig. 2D).
Associations of 4 major fertility-related factors (age, day-3 FSH, AMH, and Hishi) with achievement of cumulative live delivery were analyzed by logistic regression analysis. Among these factors, only Hishi significantly correlated with cumulative live delivery (p < 0.05, odds ratio 5.1, 95% CI 1.4 – 18.3).
Fifteen Hishi and five control patients respectively delivered 17 (8 male, 9 female) and 5 (2 male, 3 female) normal live infants, including 2 and 0 sets of twins. Considering the 13 and 5 singleton newborns in Hishi and control groups, no significant difference was evident in body weight (2845 ± 723 vs. 3023 ± 108 g; unpaired t test) or gestational age at delivery (37.5 ± 3.0 vs. 38.0 ± 1.2 weeks; Mann-Whitney U test).
No adverse events or side effects from Hishi were observed among Hishi group patients throughout the study period, nor in their fetuses or infants.
Effect of Hishi on AGE
Serum concentrations of Nω-(carboxymethyl) arginine (CMA) and Nδ-(5-hydro-5-methyl-4-imidazolone-2-yl)-ornithine (MG-H1) decreased significantly between pre-trial and intra-trial determinations in the Hishi group (n = 29: CMA, from 0.73 ± 0.19 mmol/mol Lys to 0.53 ± 0.16, p < 0.001; paired t test; and MG-H1, from 4.78 ± 1.86 mmol/mol Lys to 3.55 ± 1.25, p < 0.001; paired t test). Significant decreases did not occur in controls (n = 30: CMA, from 0.68 ± 0.17 to 0.59 ± 0.26, p = 0.09, Wilcoxon matched-pairs signed rank test; and MG-H1, from 3.41 ± 1.48 to 3.26 ± 1.79, p = 0.4; Wilcoxon matched-pairs signed rank test). The rate of increase for MG-H1, defined as [(intra-trial value minus pre-trial value) / pre-trial value] x 100%, was significantly lower in the Hishi group than in controls (-20.8 ± 28.9%, n = 29 vs. 16.9 ± 104.2, n = 30; p < 0.05, Mann-Whitney U test; Fig. 3A), meaning that MG-H1 decreased significantly more in Hishi patients. Concentrations of CMA in FF were significantly lower in Hishi patients than controls (0.16 ± 0.13 mmol/mol Lys, n = 30 vs. 0.23 ± 0.19, n = 29; p < 0.05, Mann-Whitney U test; Fig. 3B). Thus, Hishi decreased serum CMA and MG-H1, as well as FF CMA.
Serum Nε-(carboxymethyl)lysine (CML) and toxic AGE (TAGE) decreased and serum pentosidine (Pent) increased significantly in both Hishi (n = 29) and control (n = 30) groups (CML, from 0.45 ± 0.18 mmol/mol Lys to 0.38 ± 0.15, p < 0.05 in the Hishi group and from 0.50 ± 0.21 to 0.41 ± 0.23, p < 0.01 in the control group, Wilcoxon matched-pairs signed rank test; TAGE, from 10.7 ± 3.0 U/mL, n = 28, to 8.3 ± 2.8, n = 28, p < 0.001, paired t test in the Hishi group and from 9.3 ± 3.2, n = 27, to 7.6 ± 2.3, n = 27, p < 0.001, Wilcoxon matched-pairs signed rank test in the control group; and Pent, from 82.4 ± 17.2 fmol/μL to 92.2 ± 15.6, p < 0.01, paired t test in the Hishi group and from 88.3 ± 23.2 to 101.5 ± 29.3, p < 0.01, Wilcoxon matched-pairs signed rank test in the control group). No significant differences were detected between groups in rates of increase for CML, TAGE, or Pent as well as their concentrations in FF (Mann-Whitney U test for CML; unpaired t test for TAGE and Pent; Fig. 3). Thus, changes in serum CML, TAGE, and Pent observed in both groups may reflect shared interventions such as life-style modification and suppression of ovulation by the Kaufmann cycle.
MG-H1 concentrations in FF showed a significant negative correlation with numbers of follicles measuring 12 mm or larger on the hCG day, oocytes retrieved, oocytes fertilized, and day-2 embryos in the Hishi group but not controls (Table 2, Fig. 4). In other words, follicular development and oocyte quality improved in association with decreasing FF MG-H1 concentrations in Hishi patients, but no such relationship was seen in controls. Before Hishi was given, follicular development presumably was similar between groups irrespective of MG-H1 concentrations in FF, considering that group assignment of patients was random. Hishi likely improved follicular development and oocyte quality by decreasing AGE in ovarian follicles, especially in patients showing larger reductions.
Increases in serum MG-H1 from pre-trial values correlated negatively with development of retrieved oocytes to day-2 embryos in the Hishi group but not controls (Table 2, Fig. 4); the more serum MG-H1 decreased with Hishi, the more likely oocytes developed to day-2 embryos. Pre-trial serum MG-H1 concentrations correlated positively with percentage of superior day-2 embryos among all day-2 embryos in the Hishi group but not controls, suggesting that Hishi is particularly effective in improving embryo quality in patients with elevated AGE at beginning of treatment. Further, pre-trial serum concentrations of MG-H1 showed a significant negative correlation with endometrial thickness in controls but not the Hishi group (Table 2), suggesting that Hishi improved endometrial conditions by decreasing AGE. Also, pre-trial serum TAGE concentrations showed a significant positive correlation with implantation rates per freshly transferred embryo in the Hishi group (ρ = 0.40, n = 26, p < 0.05, Spearman test) but not controls (ρ = -0.08, n = 26, p = 0.71, Spearman test), suggesting that Hishi may be particularly effective in improving endometrial receptivity in patients with elevated TAGE concentrations.
Influence of Hishi extract on clinical examination and routine laboratory findings
No significant differences were evident between Hishi and control groups in pre-trial, first intra-trial, and second intra-trial examinations, including BMI, SBP, DBP, OGTT, various biochemical analyses, and hormone measurements–except for TC, which decreased significantly in the control group (from 189 ± 33 mg/dL, n = 20, to 176 ± 20, n = 20, p < 0.05, paired t test) but not in Hishi patients (initially 196 ± 21 mg/dL, n = 18, and later 191 ± 21, n = 18, p = 0.39; paired t test). This resulted in a significantly lower intra-trial TC value in controls than in the Hishi group (p < 0.05, unpaired t test).
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.