Live birth rates with and without preimplantation genetic testing: a single-center retrospective study.

We analyzed the live birth rate in patients undergoing preimplantation genetic testing (PGT) with a single frozen embryo transfer during their first cycle, comparing it to conventional in vitro fertilization (IVF). We also examined whether different PGT types, controlled ovarian hyperstimulation protocols, and endometrial preparation protocols influenced live birth rates. Results indicate that PGT holds the potential to improve live birth rates in patients undergoing a single embryo transfer.

Initially, 422 couples who met the criteria and were willing to participate in the study were screened, 230 couples underwent PGT cycles and 192 couples underwent conventional IVF cycles. In the PGT group, 11 (4.78%) couples withdrew from the study due to failure to undergo implantation or loss to follow-up, whereas, in the IVF group, 2 (1.04%) couples withdrew from the study due to loss to follow-up. Ultimately, 219 couples were included in the PGT group, of which 87 were PGT-A, 22 were PGT-M, and 110 were PGT-SR, and 190 couples were included in the IVF group (Fig.  1 ). Fig. 1 Study recruitment and results Study recruitment and results Couples in the PGT group had higher age and years of infertility than in the IVF group, while the number of retrieved oocytes was lower than in the IVF group. However, there were no significant differences between the PGT group and the IVF group in trigger day estradiol levels and the number of available blastocysts. The endometrial thickness on the day of transfer was lower in the PGT group than in the IVF group. The weeks of pregnancy were similar in both groups. During the controlled ovarian hyperstimulation phase, more women (62.56%) in the PGT group preferred the PPOS protocol, whereas more women (56.32%) in the IVF group chose the antagonist protocol. During endometrial preparation, both groups underwent down-regulation combined with hormone replacement cycle in the highest number of women (77.63% and 61.58% in the PGT and IVF groups, respectively) (Table  1 ). Table 1 Baseline characteristics of included patients PGT Conventional IVF PGT-A PGT-M PGT-SR Number of cases 219 190 87 22 110 Age of women (years) 32.90 ± 4.61 a 31.53 ± 4.15 35.71 ± 4.36 abc 32.36 ± 4.70 30.78 ± 3.52 Age of men (years) 35.08 ± 5.64 a 33.81 ± 5.48 37.76 ± 6.03 abc 34.05 ± 6.09 33.16 ± 4.26 Years of infertility (years) 3.038 ± 2.91 a 4.16 ± 2.64 2.53 ± 2.65 ab 1.81 ± 2.82 ab 3.69 ± 2.99 Controlled ovarian hyperstimulation protocols Agonist long protocol 46 64 17 4 25 Antagonist protocol 36 107 14 2 20 PPOS protocol 137 19 56 16 65 Endometrial preparation protocols 117 Down-regulation combined with hormone Replacement cycle 170 72 18 80 Hormone replacement cycle 6 2 4 0 2 Ovulation induction cycle 3 4 1 0 2 Natural ovulation cycle 40 67 10 4 26 Trigger day oestradiol level (pg/mL) 4819 ± 3166 5079 ± 2686 3868 ± 2906 ab 3940 ± 2378 5194 ± 2971 Number of retrieved oocytes 15.55 ± 7.41 a 18.97 ± 7.80 12.27 ± 6.46ab 13.08 ± 5.17 a 16.85 ± 7.05 a Number of available blastocysts 6.58 ± 3.46 5.83 ± 3.05 5.41 ± 2.60 c 7.31 ± 2.66 6.02 ± 2.62 Endometrial thickness (mm) 9.32 ± 1.71 a 10.46 ± 5.49 8.99 ± 1.67 ab 9.50 ± 1.85 9.56 ± 1.68 a Weeks of pregnancy (weeks) 37.71 ± 1.40 37.85 ± 1.18 37.50 ± 1.69 38.10 ± 1.20 37.82 ± 1.13 “a” means p  < 0.05 vs. IVF, “b” means p  < 0.05 vs. PGT-SR, “c” means p  < 0.05vs. PGT-M Baseline characteristics of included patients “a” means p  < 0.05 vs. IVF, “b” means p  < 0.05 vs. PGT-SR, “c” means p  < 0.05vs. PGT-M Due to baseline imbalances among patients, we also conducted a binary logistic regression analysis. The results revealed that patient age, duration of infertility, number of retrieved oocytes, and endometrial thickness were not significantly associated with live birth outcomes (Supplementary Table 1). The results and their corresponding odds ratios and p-values are displayed in Tables 2 and 3 . The live birth rate in the PGT group (46.12%) was significantly higher than that in the conventional IVF group (34.74%) ( p  = 0.0195, OR 0.6218, 95% CI 0.4161–0.9189). Among them, the live birth rate in the PGT-A subgroup (48.28%) was significantly higher than that in the IVF group ( p  = 0.032, OR 0.5703, 95% CI 0.3380–0.9610); and the live birth rates in the PGT-M and PGT-SR subgroups (45.45% and 44.55%, respectively), although higher than that in the IVF group, were not statistically different ( p  = 0.321, OR 0.6387, 95% CI 0.2776–1.577 and p  = 0.0922, OR 0.6626, 95% CI 0.4096–1.074, respectively); live birth rates did not differ between the three PGT subgroups. In addition, the PGT group and PGT-A subgroups (30.82% and 25.00%, respectively) had lower miscarriage rates compared to the IVF group (42.61%) ( p  = 0.0489, OR 1.666, 95% CI 1.005–2.787 and p  = 0.0251, OR 2.227, 95% CI 1.077–4.648, respectively); the PGT-M and PGT-SR subgroups (37.50% and 33.78%, respectively) had lower miscarriage rates than the IVF group, but were still not statistically different ( p  = 0.6981, OR 1.237, 95% CI 0.4491–3.697 and p  = 0.225, OR 1.455, 95% CI 0.7929–2.647, respectively); miscarriage rates did not differ between the three PGT subgroups. Biochemical, clinical and multiple pregnancy rates did not differ significantly between groups and subgroups (Table  2 and Table  3 ). Table 2 Primary and secondary outcomes PGT Conventional IVF PGT-A PGT-M PGT-SR Live birth rate % ( n ) 46.12% (101/219) a 34.74% (66/190) 48.28% (42/87) a 45.45% (10/22) 44.55% (49/110) Biochemical pregnancy rate % ( n ) 79.91% (175/219) 74.21% (141/190) 75.86% (66/87) 90.91% (20/22) 80.91% (89/110) Clinical pregnancy rate % ( n ) 66.67% (146/219) 60.53% (115/190) 64.37% (56/87) 72.73% (16/22) 67.27% (74/110) Miscarriage rate % ( n ) 30.82% (45/146) a 42.61% (49/115) 25.00% (14/56) a 37.50% (6/16) 33.78% (25/74) Multiple pregnancy rate % ( n ) 4.11% (6/146) 2.61% (3/115) 5.36% (3/56) 6.25% (1/16) 2.70% (2/74) “a” means p  < 0.05 vs. IVF Table 3 Odds ratios and p-values for comparisons between groups A PGT vs. IVF PGT-A vs. IVF PGT-M vs. IVF PGT-SR vs. IVF OR p OR p OR p OR p Live birth rate 0.6218 (0.4161–0.9189) 0.0195 0.5703 (0.3380–0.9610) 0.032 0.6387 (0.2776–1.577) 0.321 0.6626 (0.4096–1.074) 0.0922 Biochemical pregnancy rate 0.7235 (0.4594–1.164) 0.1703 0.9156 (0.5100–1.667) 0.7691 0.2878 (0.06474–1.187) 0.0828 0.6790 (0.3870–1.201) 0.1862 Clinical pregnancy rate 0.7667 (0.5104–1.149) 0.1975 0.8488 (0.5017–1.438) 0.5415 0.5750 (0.2209–1.542) 0.2649 0.7459 (0.4559–1.223) 0.2435 Miscarriage rate 1.666 (1.005–2.787) 0.0489 2.227 (1.077–4.648) 0.0251 1.237 (0.4491–3.697) 0.6981 1.455 (0.7929–2.647) 0.225 Multiple pregnancy rate 0.6250 (0.1685–2.404) 0.5094 0.4732 (0.1082–2.090) 0.3593 0.4018 (0.05726–5.539) 0.4277 0.9643 (0.1931–5.539) 0.9686 B PGT-A vs. PGT-SR PGT-M vs. PGT-SR PGT-A vs. PGT-M OR p OR p OR p Live birth rate 1.162 (0.6498–2.026) 0.602 1.037 (0.3986–2.522) 0.9376 0.8929 (0.3308–2.253) 0.8129 Biochemical pregnancy rate 1.348 (0.6824–2.665) 0.3904 2.360 (0.5759–10.82 0.259 3.182 (0.7652–14.60) 0.1223 Clinical pregnancy rate 0.8788 (0.4850–1.605) 0.6691 1.297 (0.4956–3.582) 0.6161 1.476 (0.5437–4.181) 0.4594 Miscarriage rate 0.6533 (0.3091–1.383) 0.2792 1.176 (0.3652–3.535) 0.7767 1.800 (0.5066–6.214) 0.3249 Multiple pregnancy rate 2.038 (0.4024–11.73) 0.4358 2.400 (0.1560–21.40) 0.4735 1.178 (0.08549–8.382) 0.8906 Primary and secondary outcomes “a” means p  < 0.05 vs. IVF Odds ratios and p-values for comparisons between groups We categorized all patients into three subgroups according to different controlled ovarian hyperstimulation protocols: agonist long protocol, antagonist protocol, and PPOS protocol. Similarly, we categorized all patients into four subgroups according to different endometrial preparation protocols: down-regulation combined with hormone replacement cycle, hormone replacement cycle, ovulation induction cycle and natural ovulation cycle four subgroups. We then analyzed the live birth rates among the subgroups (Table  4 , Supplementary Table 2 and Supplementary Table 3). Table 4 Live birth rate in subgroup analysis PGT Conventional IVF PGT-A PGT-M PGT-SR Controlled ovarian hyperstimulation protocols Agonist long protocol 50.00% (23/46) 40.63% (26/64) 52.94% (9/17) 50.00% (2/4) 48.00% (12/25) Antagonist protocol 52.78% (19/36) 34.58% (37/107) 42.86% (6/14) 50.00% (1/2) 60.00% (12/20) a PPOS protocol 43.07% (59/137) a * 15.79% (3/19) 48.21% (27/56) a 43.75% (7/16) 38.46% (25/65) Endometrial preparation protocols Down-regulation combined with hormone replacement cycle 47.65% (81/170) a 29.06% (34/117) 51.39% (37/72) a 50.00% (9/18) 43.75% (35/80) a Hormone replacement cycle 33.33% (2/6) 0% (0/2) 25.00% (1/4) 0% (0/0) 50.00% (1/2) Ovulation induction cycle 0% (0/3) 50.00% (2/4) 0% (0/1) 0% (0/0) 0% (0/2) Natural ovulation cycle 45.00% (18/40) 44.78% (30/67) # 40.00% (4/10) 25.00% (1/4) 50.00% (13/26) “a” means p  < 0.05 vs. IVF, “*” means p  < 0.05 vs. agonist long protocol, “#” means p  < 0.05 vs. down-regulation combined with hormone replacement cycle Live birth rate in subgroup analysis “a” means p  < 0.05 vs. IVF, “*” means p  < 0.05 vs. agonist long protocol, “#” means p  < 0.05 vs. down-regulation combined with hormone replacement cycle In the patients of down-regulation combined with hormone replacement cycle, the live birth rate was significantly higher in the PGT group (47.65%) than in the IVF group (29.06%) ( p  = 0.0016, OR 0.4501, 95% CI 0.2701–0.7443); the PGT-A and PGT-SR subgroups (51.39% and 43.75%, respectively) also had significantly higher live birth rates than the IVF group ( p  = 0.0021, OR 0.3875, 95% CI 0.2137–0.7217 and p  = 0.0338, OR 0.5267, 95% CI 0.2863–0.9575, respectively). In the patients undergoing conventional IVF, the live birth rate was significantly higher in the natural ovulation cycle (44.78%) than in the down-regulation combined with hormone replacement cycle (29.06%) ( p  = 0.0313, OR 0.5052, and 95% CI 0.2726–0.9318). The other significant results should be referred to with caution in our opinion because the sample size was too low [ 31 , 32 ].

All couples presented to Chengdu Women’s and Children’s Central Hospital from 2021.01.01 to 2023.12.31 with a diagnosis of infertility and were scheduled for their first IVF cycle. Intracytoplasmic sperm injection (ICSI) is used for all fertilization procedures. Couples with three or more good quality blastocysts will be included in the study. Exclusion criteria included known uterine anomalies (e.g., congenital malformations of the uterus; untreated uterine septum, adenomyosis, or submucosal fibroids; endometrial polyps; or intrauterine adhesions), the presence of contraindications to pregnancy, or the use of donor oocytes or sperm to achieve pregnancy. All couples provided written informed consent on the day of oocyte retrieval. The study was approved by the Ethics Committee of Chengdu Women’s and Children’s Central Hospital (2020.5 IEC-C-V2.0 2020(12)). Controlled ovarian hyperstimulation protocols mainly consist of agonist long protocol, antagonist protocol or progestin-primed ovarian stimulation (PPOS) protocol [ 37 ]. The long protocol is a gonadotropin-releasing hormone (GnRH) agonist administered on the day 2 to day 4 of menstruation or in the mid-luteal phase of the previous cycle, and gonadotropins are started after satisfactory pituitary desensitization has been achieved; the antagonist protocol is gonadotropins from day 2 to 3 of the menstrual cycle, with concomitant GnRH antagonists when the follicle reaches 14 mm in diameter or when estrogen is more than 300 pg/ml; PPOS is the administration of exogenous progesterone from days 2 to 3 of the menstrual cycle, together with gonadotropins. When at least two follicles have an average diameter of 18 mm or more, clinicians use human chorionic gonadotropin (hCG), GnRH agonists, or both, to bring about the final maturation of the oocytes, which are then retrieved by transvaginal ultrasound after about 36 h. All IVF procedures were performed using ICSI and all embryos were cultured to the blastocyst stage. According to the Gardner Morphological Criteria [ 12 ], the blastocyst morphological score is based on three components: blastocyst expansion, inner cell mass and trophectoderm development, and blastocysts rated 4BC or higher on day 5 of embryo culture are considered to be high-quality blastocysts. Patients with three or more good quality blastocysts will be included in the study. In the PGT group, high-quality blastocysts screened by morphological criteria underwent trophoblast biopsy. Select the appropriate PGT type based on patient indications. For patients undergoing PGT-M and PGT-SR, we will concurrently perform aneuploidy screening. All the embryos obtained were cryopreserved. All patients underwent a single frozen embryo transfer. The endometrial preparation protocols included down-regulation combined with hormone replacement cycles, hormone replacement cycles, ovulation induction cycles and natural ovulation cycles. An optimal blastocyst was selected for transfer in the PGT group by combining morphological criteria and biopsy results, and blastocysts were selected in the conventional IVF group only on the basis of morphological criteria. Pregnancy and live births were recorded in all patients. The control group comprised the conventional IVF group, while patients undergoing PGT-A alone were included in the PGT-A group. The primary outcome is the live birth rate of the first frozen single-embryo transfer. Secondary outcomes included biochemical pregnancy rate, clinical pregnancy rate, miscarriage rate and multiple pregnancy rate. Subgroup analyses were performed according to controlled ovarian hyperstimulation protocols and endothelial preparation protocols. Continuous baseline characteristics of patients were expressed as mean ± standard deviation (SD), and differences between groups were compared by the Wilcoxon rank sum test owing to the non-normality of the variables. Categorical variables were expressed in the form of percentages and compared by the chi-square test. Effect sizes for differences in categorical variables were reported using odds ratios (ORs), and 95% confidence intervals (CIs) for ORs were calculated using the Baptista-Pike method. Couples lost to follow-up were not included in the final statistics. Statistical significance was defined as two-sided p  < 0.05. All statistical analyses were performed using SPSS 19.0 software.

Until today, selecting the best embryos for transfer remains a major challenge for IVF, especially when clinicians choose the single embryo transfer (SET) protocol for a variety of different reasons [ 16 ]. There have been randomized controlled trials and Meta-analyses confirming the poor clinical outcomes of SET [ 13 , 18 ]; whereas multiple embryo transfer is often used to increase embryo implantation rates in clinical practice, but this regimen increases the risk of multiple pregnancies [ 2 , 10 ]. It is still widely recognized that the morphology of the embryo is closely related to its ability to survive and develop during conventional IVF [ 1 , 23 ]. Although culturing embryos to the blastocyst stage has been proven to improve clinical outcomes in SET [ 30 ], morphologically normal blastocysts are still at risk of abnormal genetic material (Forman, Hong, Franasiak, & Scott, 2014; [ 9 , 24 ]). Therefore, three PGT techniques have been introduced for assisted embryo selection with the expectation that they improve assisted reproductive outcomes [ 22 ]. In our single-center retrospective cohort study, a total of 409 couples were included in the statistics, all women underwent a single frozen embryo transfer, and we only counted the clinical outcomes of the first embryo transfer. The results showed that the live birth rate of PGT was higher than that of conventional IVF, in which the live birth rate of PGT- A was also higher than that of conventional IVF, although the live birth rates of PGT-M and PGT-SR did not differ from those of conventional IVF, but were numerically higher. In addition, PGT and PGT-A had lower miscarriage rates compared to conventional IVF. However, biochemical pregnancy rate, clinical pregnancy rate and multiple pregnancy rate did not differ between groups. Whether PGT-A can improve the live birth rate of patients is still a controversial topic until now. At first, PGT-A was thought to increase the live birth rate of embryos in a single transfer [ 35 ], reduce the risk of multiple births (Practice Committee of the American Society for Reproductive & the Practice Committee for the Society for Assisted Reproductive Technologies. Electronic address 2021), decrease the risk of early pregnancy and shorten the duration to pregnancy [ 28 ]. Early clinical trials and meta-analyses also support this view [ 6 , 27 ]. However, in recent years, more and more counter-arguments have begun to emerge. Early studies often analyzed live birth rates on the basis of a single embryo transfer when patients had high-quality blastocysts [ 28 , 34 , 35 ], and these studies are considered to have subjectively excluded from the analysis cycles with fewer embryos, poorer embryo quality, no blastocysts, and no transferable embryos, invisibly overstating the effect of PGT-A on the live birth rate [ 19 ]. More and more studies have begun to report that the cumulative live birth rate is the reasonable metric to assess the utility of PGT-A [ 7 , 19 , 44 ], and that PGT-A did not improve the cumulative live birth rate in patients [ 26 , 29 , 33 , 40 ]. In our study, both the PGT group and the conventional IVF group included only patients with excellent blastocyst quality in an attempt to reduce selection bias, but the results still favored PGT-A. There are two main possible reasons. First, because of the short period of time that PGT has been conducted in our center, it resulted in our inability to count the cumulative live birth rate, and we still used the live birth rate of a single transfer as the outcome indicator. Second, we did not transfer mosaic embryos, which may have had an impact on the final outcome. Some studies have shown that transferring mosaic embryos increases the rate of implantation failure and miscarriage despite the chance of a healthy delivery [ 4 , 21 , 25 , 41 , 42 ]. For safety reasons, no mosaic embryos were transferred in our study. There are very few studies on the effect of PGT-M and PGT-SR on live birth rates; after all, their own purpose is to avoid the transmission of abnormal genetic material, not to improve the live birth rates of patients. A few studies have shown that PGT-M and PGT-SR do not have a significant effect on the live birth rate; rather, the live birth rate of patients with PGT-M or PGT-SR is more closely related to the patient’s own age, endocrine status, oocyte quality and embryo quality [ 3 , 17 , 36 , 38 ]. In our study, due to the small number of patients in the PGT-M subgroup, the statistical results are not representative of the patient population and thus need to be considered with caution. However, it is reasonable that patients with PGT-SR did not have a significant profit in live birth rate. In our secondary outcomes, PGT-A significantly reduced the miscarriage rate in patients, which seems to be in line with the current aim of using PGT-A technique [ 28 ],Practice Committee of the American Society for Reproductive & the Practice Committee for the Society for Assisted Reproductive Technologies. Electronic address, 2021; [ 35 ]. However, biochemical pregnancy rates, clinical pregnancy rates, and multiple pregnancy rates did not differ between PGT and conventional IVF, and perhaps the effect of PGT on pregnancy is not limited to early pregnancies [ 44 ], but is also reflected in ongoing pregnancies. We also subgrouped patients by different controlled ovarian hyperstimulation protocols and endometrial preparation protocols. Our results showed that when the PPOS protocol was used, the live birth rate in the PGT group and PGT-A subgroup would be higher than that in the conventional IVF group, and among patients undergoing conventional IVF, a higher live birth rate would be obtained with the agonist long protocol than with the PPOS protocol. Unfortunately, however, too few patients in the conventional IVF group used the PPOS protocol, resulting in these results having to be taken into account very carefully. It has been reported in the literature that in fresh transplantation cycles, the agonist long protocol can indeed be advantageous in terms of live birth rate [ 20 ], and that a low GN dose is more favorable to improve the live birth rate of patients [ 14 ], but in frozen transplantation cycles, the effect of different controlled ovarian hyperstimulation protocols do not seem to have as pronounced an effect on the live birth rate [ 14 ]. The effect of the endometrial preparation protocol on the live birth rate of patients appears to be more pronounced for frozen transplantation cycles. Previous studies have shown that down-regulation combined with hormone replacement cycle and hormone replacement cycle are associated with significantly lower live birth rates and patients are at a higher risk of gestational hypertension, postpartum hemorrhage, and preterm labor; in contrast, the natural ovulation cycle resulted in optimal pregnancy outcomes [ 43 ]. This is also consistent with our results, but the number of patients who chose the ovulation induction cycle was too small for the current study to give constructive conclusions. However, it is interesting to note that in the group of patients who chose down-regulation combined with hormone replacement cycle, the live birth rate of PGT was significantly higher than that of the conventional IVF group, which has not been reported before. We believe that this is likely to be related to male factors. There are certain limitations to our study. The first and foremost is the insufficient sample size. The small sample size resulted in a restrictive interpretation of many of the results in the subgroup analyses. Similarly, because of the small sample size, this study did not further differentiate between age groups for subgroup analysis. Second, due to the relatively short period of time that PGT has been conducted in our center, the cumulative live birth rate was not adopted as the primary outcome in this study, which may have potentially inflated the contribution of PGT to the live birth rate. In addition, the baseline characteristics of the patients were not consistent, and it is good to note that we do not believe this had a significant impact on our results. These issues are expected to be improved in future large studies.

In vitro fertilization (IVF) is an important assisted reproductive technology (ART), and the transfer of high-quality embryos in receptive endometrium is crucial for IVF [ 16 ]. Selection of the best embryos for transfer is a major challenge in the IVF process, especially in single embryo transfer strategies. The most common method of embryo selection is still based on morphological grading criteria [ 1 , 23 ]. However, there has been substantial evidence that abnormalities in embryonic genetic material are also a significant cause of IVF failure, and therefore preimplantation genetic testing (PGT) has been employed in the process of embryo selection [ 3 ], Committee et al. 2020; Practice Committees of the American Society for Reproductive, the Society for Assisted Reproductive Technology. Electronic address, Practice Committees of the American Society for Reproductive, & the Society for Assisted Reproductive 2018). PGT is a diagnostic procedure that analyses the embryonic genetic material in order to reduce the risk of abnormalities of the genetic material and to increase the likelihood of a successful pregnancy [ 15 ]. Until today, PGT has been developed in three ways: preimplantation genetic testing for aneuploidy (PGT-A), preimplantation genetic testing for monogenic disorders (PGT-M), and preimplantation genetic testing for structural rearrangements (PGT-SR), in order to improve the selection of embryos in different situations [ 5 ]. As the combination of morphological criteria and PGT evaluates the embryos to be transferred from different latitudes, it has led to an elevated likelihood of successful pregnancies and a reduction in spontaneous abortions and genetically abnormal fetuses. However, it has been shown that invasive PGT manipulations may have a slight impact on embryo viability, and limitations in testing techniques can lead to false positives or false negatives [ 15 ]; PGT-A may miss the diagnosis of mosaic embryos and incorrectly assess the embryo’s reproductive potential [ 41 , 42 ]; PGT-M is unable to diagnose mutations that are not recorded in the current analysis strategy [ 39 ]; and the variability of chromosomal rearrangements significantly influences the complexity and accuracy of PGT-SR [ 11 ]. Each of these challenges has the potential to have an impact on the final assisted reproduction treatment outcome. In order to assess the impact of PGT on IVF treatment outcomes in different patients, as well as to provide additional clinical data for PGT prognostic studies, this study reviewed the live birth rate of IVF combining morphological criteria and PGT at our center, and compared it with the live birth rate of conventional IVF based on morphological criteria only [ 8 ].

Below is the link to the electronic supplementary material. Supplementary file1 (DOCX 26 KB) Supplementary file1 (DOCX 26 KB)