Hum. Reprod. Advance Access published March 6, 2015 Human Reproduction, Vol.0, No.0 pp. 1 –7, 2015 doi:10.1093/humrep/dev049

ORIGINAL ARTICLE Embryology

Effect of embryo culture media on percentage of males at birth Jinliang Zhu 1,2,3, Xinjie Zhuang 1,2,3, Lixue Chen1,2,3, Ping Liu 1,2,3,*, and Jie Qiao 1,2,3,* 1

Department of Obstetrics and Gynecology, Reproductive Medical Center, Peking University Third Hospital, Beijing, China 2Key Laboratory of Assisted Reproduction, Ministry of Education, Beijing, China 3Beijing Key Laboratory of Reproductive Endocrinology and Assisted Reproduction, Beijing, China

Submitted on September 21, 2014; resubmitted on February 6, 2015; accepted on February 11, 2015

study question: Does embryo culture medium influence the percentage of males at birth? summary answer: The percentage of males delivered after ICSI cycles using G5TM medium was statistically significantly higher than after cycles where Global, G5TM PLUS, and Quinn’s Advantage Media were used.

what is known already: Male and female embryos have different physiologies during preimplantation development. Manipulating the energy substrate and adding growth factors have a differential impact on the development of male and female embryos. study design, size and duration: This was a retrospective analysis of the percentage of males at birth, and included 4411 singletons born from fresh embryo transfer cycles between January 2011 and August 2013 at the Center for Reproductive Medicine of Third Hospital Peking University. participants/materials, setting, and methods: Only singleton gestations were included. Participants were excluded if preimplantation genetic diagnosis, donor oocytes and donor sperm were used. The database between January 2011 and August 2013 was searched with unique medical record number, all patients were present in the database with only one cycle. Demographics, cycle characteristics and the percentage of male babies in the four culture media groups were compared with analysis of variance or x 2 tests. Multivariable logistic regression was done to determine the association between the sex at birth and culture media after adjusting for other confounding factors, including parental age, parental BMI, type of infertility, parity, number of embryos transferred, number of early gestational sacs, cycles with testicular sperm aspiration (TESA)/percutaneous epididymal sperm aspiration (PESA)/testicular sperm extraction (TESE), number of oocytes retrieved, cycles with blastocyst transfers, and gestational age within ICSI group. main results and the role of chance: Within the IVF group, the percentage of males at birth for G5TM , Global, Quinn’s and G5TM PLUS media were comparable (P . 0.05); however, within the ICSI group, the percentage of male babies in cycles using G5TM (56.1%) was statistically significantly higher than in cycles that used Global (47.2%; P ¼ 0.003), G5TM PLUS (47.7%; P ¼ 0.005) or Quinn’s media (45.0%; P ¼ 0.009). There were no statistically significant differences in the percentage of males at birth between cycles that used Global, G5TM PLUS and Quinn’s media (P . 0.05). Multivariable logistic regression indicated that culture media (G5TM versus Global, G5TM PLUS, and Quinn’s) were significantly associated with the sex at birth (P ¼ 0.008) after adjusting for parental age, parental BMI, type of infertility, parity, number of embryos transferred, number of early gestational sacs, cycles with TESA/PESA/TESE, number of oocytes retrieved, cycles with blastocyst transfers, and gestational age.

limitations and reasons for caution: This study was not a randomized controlled trial and allocation of treatment cycles over the four media was not completely at random. Cigarette smoking was not included in the current study because this confounding factor was not registered in our database. Moreover, intra-variability of sperm selection between the five embryologists may directly affect the percentage of males.

wider implications of these findings: Our study suggests that human embryogenesis responds differently to G5TM , Global, G5TM PLUS and Quinn’s Advantage Medium. This finding can be generalized to other commercial culture media.

study funding/competing interest(s): National Natural Science Foundation of China for Young Scholars (81300483 and 81200466). The authors have no conflicts of interest to declare. & The Author 2015. Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved. For Permissions, please email: [email protected]

Downloaded from http://humrep.oxfordjournals.org/ at University of Aberdeen on April 15, 2015

*Correspondence address. Department of Obstetrics and Gynecology, Reproductive Medical Center, Peking University Third Hospital, No. 49 North Huayuan Road, Haidian District, Beijing 100191, China. Tel: +86-10-82266849 (P.L.)/+86-10-82265080 (J.Q.); E-mail: [email protected] (P.L.)/[email protected] (J.Q.)

2

Zhu et al.

trial registration number: Not applicable. Key words: sex at birth / culture medium / singleton gestation / ICSI / IVF

Introduction

Materials and Methods Participants and stimulation protocol This was a retrospective analysis of the percentage of males at birth, and included 4411 singleton gestations born from fresh embryo transfer between January 2011 and August 2013 at the Center for Reproductive Medicine of Third Hospital Peking University. Only singleton gestations were included. Participants were excluded if preimplantation genetic diagnosis, donor oocytes and donor sperm were used. The database between January 2011 and August 2013 was searched with unique medical record number; all patients were present in the database with only one cycle. Women underwent controlled ovarian hyper-stimulation with a gonadotrophin-releasing hormone agonist or antagonist protocol. Ovarian follicle development was monitored based on serum estradiol levels and transvaginal ultrasonographic measurements. When at least one follicle reached a mean diameter of 18 mm and the estradiol concentration exceeded 500 pg/ml, human chorionic gonadotrophin (Serono, Aubonne, Switzerland) was administered before ultrasonography-guided oocyte

Laboratory protocol The following four commercially available culture media were used: G5TM (Vitrolife, Gothenburg, Sweden); Global (IVF Online, Toronto, Canada); Quinn’s Advantage Medium (SAGE, Pasadena, CA, USA); and G5TM PLUS (Vitrolife, Gothenburg, Sweden). The corresponding protein sources used to supplement the media were as follows: human serum albumin (has)-solutionTM (Vitrolife); HSA solution (IVF Online); and Quinn’s Advantage Serum Protein Substitute (SPS; SAGE). For fertilization media, 7 ml HSA, 10 ml HSA or 6 ml SPS were added to Vitrolife (60 ml), Global (100 ml) or Quinn’s media (50 ml), respectively, the final protein concentration being 10 mg/ml, 9 mg/ml and 5 mg/ml in basal medium, respectively; For embryo cleavage media, 1.6 ml HSA, 5 ml HSA and 6 ml SPS were added to Vitrolife (30 ml), Global (50 ml) and Quinn’s media (50 ml), respectively, the final concentration being 5 mg/ml, 9 mg/ml and 5 mg/ml in basal media, respectively. G5TM plus protein was ready-to-use from the supplier. Quinn’s Advantage Medium (SAGE) was discontinued in July 2012, and G5TM PLUS was used from January 2011. In our center, one type of medium is typically used for 3 consecutive days, then changed to another medium. Mineral oil was obtained from Sigma (St. Louis, MO, USA) and used after washing and sterile filtration. IVF and ICSI were performed according to the routine laboratory insemination procedures on the day of oocyte retrieval. In the case of IVF, oocytes were inseminated in fertilization medium. After checking for the presence of two pro-nuclei and two polar bodies, zygotes were transferred to cleavage stage medium. In the case of ICSI, oocytes were cultured in cleavage stage medium after injection. Embryo morphology was evaluated 68 – 72 h after insemination with respect to cell number and fragmentation. The number of embryos transferred was determined based on patient age, number of IVF cycles and embryo quality. From April 2013, a maximum of two embryos were transferred into the uterus on Day 3.

Statistical analysis All statistical analyses were performed with the Statistical Package for the Social Sciences software (SPSS, version 17.0; SPSS, Inc., Chicago, IL, USA). Continuous variables were compared using analysis of variance, and categorical variables were evaluated with x 2 tests. LSD multiple comparisons were used to evaluate difference between groups. G5 group was defined as the reference group; other three groups were compared with reference group in this study. All tests were two-sided, and a P-value ,0.05 was considered statistically significant. Multivariable logistic regression analyses were used to evaluate the possible relationship between culture media and sex at birth after adjusting for other potential confounding factors, including parental age, parental BMI, type of infertility, parity, cycles with TESA/PESA/TESE, number of oocytes retrieved, cycles with blastocyst transfers, the number of embryos transferred, the number of early gestational sacs and gestational age within ICSI group. Culture media was defined as a categorical variable, and other confounding factors were defined as continuous variables (parental age, parental BMI, number of oocytes retrieved and gestational age) or binary variables (type of infertility, parity, cycles with TESA/PESA/TESE, cycles with blastocyst transfers, the number of early gestational sacs, the number of embryos transferred and sex at birth). The results of multivariable logistic regression indicated the effect of each independent factor on sex at

Downloaded from http://humrep.oxfordjournals.org/ at University of Aberdeen on April 15, 2015

In Western countries the percentage of males is 51.5% at birth (Hesketh and Xing, 2006). There is a loss of male predominance as the population ages, reaching equilibrium by the third and fourth decades of life, with a further decline in older age. The percentage of males at birth is 54.3% in mainland China (China World Databank, February 2012). It has been established that the sex at birth is skewed to fewer boys by many factors, including parental age .40 years of age (Jacobsen et al., 1999; Nicolich et al., 2000), multiparas (Orvos et al., 2001), parental smoking (Fukuda et al., 2002; Parazzini et al., 2005), paternal exposure to dioxin (Mocarelli et al., 2000; Ryan et al., 2002) and stressors (e.g. earthquake and economic depression; Fukuda et al., 1998; Catalano, 2003). In assisted reproduction technology, Luke et al. (2009) reported that ICSI reduced the proportion of male newborns compared with IVF, and Maalouf et al. (2014) reported that 3% more male infants were born from IVF than ICSI. In contrast, blastocyst transfer skewed the sex at birth toward males compared with cleavage stage embryo transfer because male embryos develop faster than their female counterparts, with 12% more male newborns from blastocyst transfer than cleavage transfer (Luna et al., 2007). Other studies, however, have not shown a relationship between blastocyst transfer and sex bias in offspring (Weston et al., 2009; Lin et al., 2010). In animal studies, manipulating energy substrates and adding growth factors have been reported to impact development and the sex of blastocysts (Gutierrez-Adan et al., 2001; Kimura et al., 2005, 2008; Rubessa et al., 2011). The above data demonstrate that the sex ratio at birth is affected by in vitro embryo culture and assisted reproduction technologies. Nevertheless, the effect of culture media on the human sex ratio at birth has not been investigated. The aim of this study was to determine whether four different culture media (G5TM , Global, G5TM PLUS and Quinn’s) influence the percentage of males at birth in singleton gestations.

retrieval. Luteal support was initiated on the day after oocyte retrieval using 60 mg of progesterone (Xianju Pharmacy, Zhejiang, China).

3

Culture media and percentage of males at birth

birth, which ruled out the possibility that other confounding factors influence the sex at birth.

Results

Discussion It is well known that culture medium is an important factor in determining the IVF success rate. A systematic review of randomized controlled trials regarding the effect of culture medium on the success of IVF suggests significant differences among widely used commercial culture media (Mantikou et al., 2013). With respect to clinical outcomes, including implantation, pregnancy and live birth rates, our quality control indicated that there were no significant differences between G5TM , Global and G5TM -PLUS media. Singletons resulting from G5TM , Global and Quinn’s media have been reported to have similar absolute birthweights (Lin et al., 2013). Our recent study indicated that singletons born from embryos cultured in G5TM -PLUS had a greater gestational age- and gender-adjusted birthweight (Z scores) than embryos cultured in G5TM ; we also showed that more large for gestational age infants were in the G5TM -PLUS group compared with the G5TM group (Zhu et al., 2014). Although both G5TM and G5TM -PLUS are produced by the same manufacturer, we speculate that the difference between the two media may be related to composition and amount of protein from the same manufacturer. The protein content and amount of G5TM , Global, and Quinn’s media could be different because HSA solution was obtained from a different manufacturer. A recent study reported that a large variety of non-declared proteins were detected in commercial embryo culture media, and amount of proteins varied from batch to batch (Dyrlund et al., 2014). In light of the different clinical and neonatal outcomes with the four different culture media, we have investigated the effect of culture media on the sex ratio at birth. Previous animal studies have shown that the addition of energy substrates (e.g. glucose and glucosamine) and growth factors (e.g. myo-inositol) to culture media has an effect on the sex ratio of embryos produced in vitro (Gutierrez-Adan et al., 2001; Kimura et al., 2005, 2008; Rubessa et al., 2011; Torner et al., 2014). Little is known, however, regarding the effect of culture media on the percentage of males at birth in humans. The current study was the first to report that the choice of culture medium causes a sex bias in the offspring of couples undergoing ICSI; the sex at birth of singletons resulting from G5TM was skewed toward males, while the sex at birth from Global, G5TM -PLUS and Quinn’s media were skewed toward females within the ICSI group. It has been established that some biological factors affect the sex ratio at birth in couples undergoing assisted reproduction technology (e.g. maternal age, paternal age and parity). It is controversial whether maternal age affects the sex ratio (Jacobsen et al., 1999; Orvos et al., 2001). A recent study with a larger cohort showed that the sex ratio was not associated with advanced maternal age, but for a subgroup with pre-eclampsia, older maternal age led to a low ratio of male-to-female offspring (Rueness et al., 2012). In the current study, the demographic characteristics in the ICSI groups, including parental age, paternal BMI, parity, and type of infertility, cultured in G5TM medium did not differ from Global, Quinn’s and G5TM -PLUS media. Moreover, no difference was e21 2found in total dose of gonadotrophin, days of gonadotrophin

Downloaded from http://humrep.oxfordjournals.org/ at University of Aberdeen on April 15, 2015

The study population included 2469 and 1942 singletons following IVF and ICSI, respectively. In the ICSI group, 587, 580, 583 and 189 singletons resulted from embryos cultured in G5TM , Global, G5TM PLUS and Quinn’s, respectively. In the IVF group, 874, 643, 768 and 184 singletons resulted from embryos cultured in G5TM , Global, G5TM PLUS and Quinn’s, respectively. Each patient contributed only one singleton in this study. The demographic and cycle characteristics are shown in Table I. G5TM was defined as the reference group; Global, G5TM PLUS and Quinn’s groups were compared with the reference group. There was no statistical difference for some demographic characteristics for ICSI groups (including parental age, paternal BMI, type of infertility and parity) in the G5TM compared with the Global, G5TM PLUS and Quinn’s groups (P . 0.05). Similarly, the cycle characteristics for ICSI groups (including dose of gonadotrophin, days of gonadotrophin administration, number of oocytes retrieved, cycles with blastocyst transfers and cycles with TESA/PESA/TESE) did not differ significantly between the G5TM and the Global, G5TM PLUS, and Quinn’s groups, respectively (P . 0.05). Moreover, no significant differences were found with respect to preterm delivery and gestational age (P . 0.05); however, the maternal BMI in ICSI cycles was significantly less in Global group (P ¼ 0.025). Similarly the Quinn’s groups in ICSI cycles had significantly more embryos transferred (P ¼ 0.000). Since it was known that presence of an early second gestational sac that resulted in a vanishing twin can influence perinatal parameters of the surviving fetus, the number of cases with one, two and three embryos transferred were shown in Supplementary Table SI. As shown in Table II, ICSI significantly reduced the proportion of males compared with IVF (49.8 versus 55.1%; P , 0.001). The percentage of males of singletons conceived by ICSI resulting from ejaculated sperm was not different from TESA/PESA/TESE (50.7 versus 46.1%, P ¼ 0.110). Considering both IVF and ICSI, the percentage of males of singletons resulting from Day 3 transfer was comparable with blastocyst transfers (52.7 versus 55.8%, P ¼ 0.502). G5TM significantly skewed the sex toward males compared with the Global group (55.5 versus 50.0%; P ¼ 0.005); however, no statistical difference existed between G5TM and Quinn’s or G5TM PLUS (P . 0.05). Because the fertilization method significantly influenced the percentage of males at birth, the singletons resulting from the four different culture media were categorized as IVF and ICSI. Within the IVF group, there was no statistical difference in G5TM compared with the other three culture media (P . 0.05); the percentage of males of singletons resulting from Global, Quinn’s and G5TM PLUS was also comparable (P . 0.05). In the ICSI group, however, compared with the G5TM group (56.1% males) there was a statistically significantly lower percentage of males in the Global (47.2%; P ¼ 0.003), G5TM PLUS (47.7%; P ¼ 0.005) and Quinn’s groups (45.0%; P ¼ 0.009). No statistically significant difference was demonstrated between the Global, G5TM PLUS and Quinn’s groups (P . 0.05). Finally, multivariable logistic regression was performed to determine the association between the sex at birth and culture medium with other confounding factors, including parental age, parental BMI, type of infertility, parity, number of embryos transferred, number of early gestational sacs,

cycles with TESA/PESA/TESE, number of oocytes retrieved, cycles with blastocyst transfers, and gestational age within the ICSI group. As shown in Table III, the sex at birth was statistically significantly associated with paternal BMI (P ¼ 0.034), gestational age (P ¼ 0.003) and culture medium (G5TM versus Global, P ¼ 0.003; G5TM versus G5TM PLUS; P ¼ 0.008; and G5TM versus Quinn’s; P ¼ 0.015).

4

Table I Demographic and cycle characteristics data are presented as % or mean + SD. IVF

ICSI

............................................................................................ ............................................................................................. IVF cycles (n 5 2469)

TM

G5 (n 5 874)

Global (n 5 643)

TM

G5 PLUS (n 5 768)

Quinn’s (n 5 184)

ICSI cycles (n 5 1942)

G5TM (n 5 587)

Global (n 5 580)

G5TM PLUS (n 5 583)

Quinn’s (n 5 189)

.......................................................................................................................................................................................................................................................... Maternal age (years)

32.4 + 4.08

32.2 + 4.07

32.4 + 4.18

32.4 + 4.10

33.0 + 3.67

31.6 + 4.48

31.5 + 4.47

31.9 + 4.64

31.3 + 4.28

P-value

–

–

0.494

0.349

0.019

–

–

0.150

0.409

31.7 + 4.54 0.550

Paternal age (years)

34.0 + 4.84

33.8 + 4.88

34.0 + 4.82

33.9 + 4.84

34.5 + 4.71

33.4 + 5.57

33.5 + 5.65

33.7 + 5.84

33.0 + 5.25

33.5 + 5.36

P-value

–

–

0.374

0.730

0.108

–

–

0.517

0.107

0.956

Maternal BMI (kg/m2)

22.5 + 3.32

22.5 + 3.33

22.4 + 3.28

22.4 + 3.33

22.5 + 3.33

22.3 + 3.35

22.5 + 3.74

22.0 + 3.00

22.4 + 3.40

22.1 + 2.94

P-value

–

–

0.482

0.618

0.926

–

–

0.025

0.717

0.162

Paternal BMI (kg/m2)

25.2 + 3.49

25.3 + 3.35

25.0 + 3.31

25.3 + 3.80

25.4 + 3.40

25.2 + 3.77

25.3 + 3.92

25.4 + 3.65

25.1 + 3.83

25.0 + 3.48

P-value

–

–

0.185

0.853

0.586

–

–

0.494

0.606

0.349

Primary infertility (%)

43.3

44.2

40.3

44.0

46.7

67.9

68.0

66.8

69.1

67.2

P-value

–

–

0.141

0.960

0.568

–

–

0.708

0.706

0.858

Woman’s first child (%)

92.3

91.7

92.5

93.1

90.8

95.4

95.7

95.2

95.0

90.8

0.566

0.307

0.664

–

–

0.675

0.581

P-value Dose of gonadotrophin (IU)

2996 + 1341 3027 + 1349 2957 + 1373 3020 + 1324

P-value

–

Duration of gonadotrophin administration (days) 11.8 + 2.45

2883 + 1253 2808 + 1316

2838 + 1258 2783 + 1369 2784 + 1278

0.836 2863 + 1444

–

0.318

0.916

0.186

–

–

0.482

0.488

0.820

11.9 + 2.42

11.6 + 2.45

11.8 + 2.47

11.7 + 2.44

11.6 + 2.44

11.6 + 2.49

11.5 + 2.29

11.6 + 2.54

11.5 + 2.42

P-value

–

–

0.034

0.573

0.198

–

–

0.557

0.678

0.464

Number of oocytes retrieved

12.5 + 6.12

12.5 + 6.06

12.2 + 5.98

12.7 + 6.30

12.5 + 6.08

12.9 + 6.39

12.8 + 6.14

12.8 + 6.32

12.9 + 6.67

13.3 + 6.50

P-value

–

–

0.432

0.482

0.930

–

–

0.948

0.977

0.403

Number of embryos transferred

2.2 + 0.49

2.2 + 0.47

2.2 + 0.50

2.2 + 0.49

2.3 + 0.59

2.2 + 0.51

2.2 + 0.51

2.2 + 0.49

2.2 + 0.49

2.4 + 0.54

P-value

–

–

0.315

0.006

0.006

–

–

0.19

0.175

,0.001

Cycles with blastocyst transfer (%)

3.3

4.1

3.0

2.6

3.3

3.4

2.9

3.5

3.6

4.2

P-value

–

–

0.267

0.102

0.683

–

–

0.620

0.515

0.351

Cycles with TESA/PESA/TESE (%)

–

–

–

–

–

19.7

18.9

17.8

20.9%

24.3

P-value

–

–

–

–

–

–

–

0.650

0.421

0.118

Preterm delivery N (%)

198(8.0)

76(8.7)

53(8.2)

59(7.7)

10(5.4)

130(6.7)

50(8.5)

37(6.4)

32(5.5)

11(5.8)

P-value

–

–

0.781

0.473

0.181

–

–

0.182

0.051

0.278

Gestational age (weeks)

38.5 + 1.75

38.4 + 1.73

38.5 + 1.51

38.5 + 1.71

38.6 + 1.85

38.6 + 1.78

38.5 + 1.71

38.6 + 1.63

38.6 + 2.07

38.6 + 1.49

P-value

–

–

0.331

0.762

0.289

–

–

0.300

0.166

0.487

Continuous variables were compared using analysis of variance and categorical variables were evaluated with chi-squared tests. LSD multiple comparison was used to evaluate differences between groups; G5TM was defined as the reference group. P-values indicate statistical difference between Global, G5TM PLUS, and Quinn’s Advantage Medium, respectively, versus G5TM .

Zhu et al.

Downloaded from http://humrep.oxfordjournals.org/ at University of Aberdeen on April 15, 2015

5

Culture media and percentage of males at birth

Table II Percentage of males based on different culture media and cycle characteristics. Insemination method

Cycle characteristic or medium

Singletons (n)

Male (n)

Male percent

P-value*

............................................................................................................................................................................................. ,0.001

IVF

All

2469

1361

55.1

ICSI

All

1941

967

49.8

ICSI

Ejaculated sperm

1559

791

50.7

ICSI

TESA/PESA/TESE

46.1

IVF + ICSI

Days 3 transfer

IVF + ICSI

Blastocyst transfer

IVF + ICSI

G5TM

IVF + ICSI

Global

1224

612

50.0

0.005

IVF + ICSI

G5TM PLUS

1353

708

52.3

0.096

IVF + ICSI

Quinn’s

373

197

52.8

0.352

IVF

G5TM

874

482

55.2

382

176

4263

2246

52.7

147

82

55.8

1461

811

55.5

0.110 0.502

Global

643

338

52.6

0.322

G5TM PLUS

768

429

55.9

0.804

IVF

Quinn’s

184

112

60.9

0.165

ICSI

G5TM

587

329

56.1

ICSI

Global

581

274

47.2

0.003

ICSI

G5TM PLUS

585

279

47.7

0.005

ICSI

Quinn’s

189

85

45.0

0.009

*Chi-square test; for comparisons of media all P-values refer to differences from G5TM .

Table III Multivariable logistic regression analysis of factors associated with the sex of 1942 ICSI singletons. OR

95% CI

P-value

........................................................................................ Maternal age (per year)

1.00 0.97–1.04 0.873

Paternal age (per year)

1.00 0.98–1.03 0.914

Maternal BMI (per kg/m2)

1.02 0.99–1.05 0.292

Paternal BMI (per kg/m2)

1.03 1.00–1.05 0.034

Parity (higher order versus first child)

0.85 0.53–1.37 0.509

Type of infertility (secondary versus primary)

1.05 0.84–1.30 0.694

Sperm origin (TESA/PESA/TESE versus ejaculated)

0.82 0.64–1.04 0.103

Number of oocytes retrieved (per oocyte) 1.02 1.00–1.03 0.068 Number of embryos transferred (per embryo)

0.99 0.62–1.58 0.967

Number of gestational sac (1 versus 2 or 3) 0.94 0.65–1.36 0.756 Day 3 versus blastocyst transfer

1.19 0.70–2.02 0.525

Gestational age (per week)

0.91 0.86–0.97 0.003

Culture media (G5TM [ref])

0.008§

Culture media (1) [Global]

0.70 0.55–0.89 0.003

TM

Culture media (2) [G5

PLUS]

Culture media (3) [Quinn’s]

0.72 0.57–0.92 0.008 0.66 0.47–0.92 0.015

Culture media was defined as a categorical variable. § Overall significance of a culture medium effect.

administration, number of oocytes retrieved, cycles with blastocyst transfers, and cycles with TESA/PESA/TESE between G5TM and Global, Quinn’s, and G5TM -PLUS media, respectively. However, more preterm deliveries (,37 weeks) occurred in the G5TM group, although the difference did not reach statistical significance. In our study population, preterm births have more male newborns than term births (58 versus 52%). Gestational age is the main confounding factor, which could distort the conclusions in this study, therefore gestational age and other potential confounding factors were included in multivariable logistic regression and adjusted. The results indicated culture media is still associated with sex at birth after adjusting for other confounding factors. Sex bias in relation to assisted reproduction technology has been extensively investigated. It has been reported that more male newborns are delivered from blastocyst transfers than cleavage transfers because male embryos may develop faster than their female counterparts (Luna et al., 2007). A higher percentage of blastocyst transfers does not reflect a bias in Global, G5TM -PLUS and Quinn’s media as blastocyst transfer is associated with more male bias. Previous large cohort studies have proposed that ICSI skews the sex ratio toward females compared with IVF, although the biological mechanism is not known (Luke et al., 2009; Maalouf et al., 2014). In the current study, our results are in agreement with previous cohort studies; specifically, ICSI significantly reduced the percentage of males compared with IVF (5% more males in IVF than ICSI). In the current study, different culture media yielded a different percentage of males at birth within ICSI cycles (an approximate 9% greater percentage of males with G5TM media than Global, G5TM -PLUS, and Quinn’s media). The different composition of basal culture media and

Downloaded from http://humrep.oxfordjournals.org/ at University of Aberdeen on April 15, 2015

IVF IVF

6

(×200) (Setti et al., 2012). Therefore, selecting sperm with good morphology under high magnification can bias the sex toward females; however, ICSI procedures were performed by five embryologists under the same ×200 magnification instead of IMSI in our center. Moreover, one type of medium is typically used for three consecutive days, then changed to another medium, thus guaranteeing that embryologists are allocated to different culture media groups with no intervention. This design in culture media use minimizes the effect of different embryologists on the sex bias in the different four culture media groups. Of note, a major limitation of this study was that patients were not randomly allocated to the treatment groups since this was an observational study. In this study, we suggested that culture medium is associated with sex bias at birth after ICSI. Because it is very difficult to define the normal ratio of males-to-females, which culture medium is optimal for human embryogenesis is still not known. Specific components and concentrations of culture media by far are not known, which is becoming more and more obvious in the scientific community. Therefore, disclosure of the ingredients of culture medium should be great importance to users to make choices about which culture medium to choose. In addition, the ranges of ICSI application cannot be expanded arbitrarily; ICSI should be used under clear medical indications. It is very difficult to predict future implications using a single center database involving sex bias. Since ART related and other unknown factors may contribute to sex bias, the degree to which culture media influences on sex bias needs to be further assessed with a nationwide database. In conclusion, we found that embryo culture medium modifies the percentage of males of singleton gestations at birth. This was due to preferential loss of male or female embryos exposed to different in vitro culture conditions; thus transfer of good quality embryos exhibits sex bias in different culture media. At present, children born following assisted reproductive technology (ART) account for only 2–3% of all births, therefore such a small percentage of sex bias may be neglected; however, effect of culture media on sex bias should be investigated further with a nationwide database if the percentage of ART births continues to rise.

Supplementary data Supplementary data are available at http://humrep.oxfordjournals.org/.

Authors’ roles Conceived and designed the study: P.L., J.Q.; coordinated data collection: J.L.Z., X.J.Z. and L.X.C.; analyzed the data: J.L.Z. and L.X.C.; drafted the manuscript: J.L.Z. All authors interpreted the data.

Funding This study was funded by the National Natural Science Foundation of China for Young Scholars (81300483 and 81200466).

Conflict of interest None declared.

Downloaded from http://humrep.oxfordjournals.org/ at University of Aberdeen on April 15, 2015

proteins between G5TM , Global, G5TM -PLUS, and Quinn’s may account for the sex bias of singletons at birth. Male and female embryos respond differently to in vitro culture media conditions (Gutierrez-Adan et al., 2000). The main difference between male and female embryos is the presence of an additional X chromosome in female embryos compared with male embryos. Double-dose activity of genes controlling glucose metabolism (glucose-6-phosphate dehydrogenase) and antioxidants (hypoxanthine phosphoribosyl transferase [HPRT]) located on the X chromosome lead to energy metabolism and growth rate differences with respect to male and female embryos (Tiffin et al., 1991; Peippo and Bredbacka, 1995; Gutierrez-Adan et al., 2000). Growth rate differences related to embryo sex have been observed in vivo on Day 3 and in vitro extended culture (blastocyst culture from Day 3), as evidenced by female embryos growing faster than male embryos in vivo before Day 3; however, in vitro extended culture supports male embryo development more than female embryo development (Peippo and Bredbacka, 1995). In light of above findings, we speculate that human embryos may react differently to G5TM , Global, G5TM -PLUS and Quinn’s culture media in the early days of in vitro culture. Nevertheless, sex bias at birth was not observed in G5TM , Global, G5TM -PLUS and Quinn’s culture media within IVF cycles. Genetically diagnosed IVF 4–8 cell embryos have been reported to exhibit an 60% male to 40% female ratio (Griffin et al., 1994); the ICSI counterpart embryos exhibit an 52% male to 48% female ratio (Tarin et al., 2014). Although this number may be biased by a relatively small number of embryos, the percentage of males of ICSI embryos was closer to 50% than IVF embryos. This finding was supported by a retrospective analysis with a large number of samples, in which the percentage of males at birth of ICSI and IVF was 50 and 53%, respectively (Dean et al., 2010; Maalouf et al., 2014). In addition, ICSI embryos are simultaneously fertilized relative to IVF embryos. Therefore, it is thought that the sex ratio imbalance of human embryos at fertilization (a greater percentage of male embryos) and the asynchronous fertilization time in IVF may mask the effect of culture media on the growth rate of human embryos during the first 3 days. To adjust for the effect of confounding factors on sex in the current study, we performed multivariable logistic regression with data derived from ICSI singleton gestations, which showed that paternal BMI, length of gestation and culture media (G5TM versus Global, G5TM PLUS, and Quinn’s) were significantly associated with sex bias. This result is consistent with a previous study in which there was a relationship between preterm birth and sex bias (Vatten and Skjaerven, 2004). Extremely skewed sex ratio was found in Quinns’ group for IVF cycles (male percentage in Quinns’ medium: 60.9%); therefore it is speculated that different fertilization media used in this study possibly confound skewed sex ratio observed in IVF group. Concentration and type of proteins (HSA versus SPS) supplemented in fertilization media might directly influence motility, viability and fertilization capability of sperm. However, observational study cannot interpret effect of fertilization media on extremely skewed sex ratio at fertilization in IVF group. We should exercise caution with the main findings in this study because not all confounding factors were adjusted. Smoking cigarettes were not included, as this factor was not registered in our dataset; more importantly, the sex ratio of singletons resulting from ICSI could be influenced by the sperm selection process of different embryologists. A randomized controlled trial has suggested that more female embryos are derived from intracytoplasmic morphologically selected sperm injection (IMSI) under high magnification (×6000) compared with ICSI

Zhu et al.

Culture media and percentage of males at birth

References

Maalouf WE, Mincheva MN, Campbell BK, Hardy IC. Effects of assisted reproductive technologies on human sex ratio at birth. Fertil Steril 2014; 101:1321– 1325. Mantikou E, Youssef MA, van Wely M, van der Veen F, Al-Inany HG, Repping S, Mastenbroek S. Embryo culture media and IVF/ICSI success rates: a systematic review. Hum Reprod Update 2013;19:210 – 220. Mocarelli P, Gerthoux PM, Ferrari E, Patterson DG Jr, Kieszak SM, Brambilla P, Vincoli N, Signorini S, Tramacere P, Carreri V et al. Paternal concentrations of dioxin and sex ratio of offspring. Lancet 2000; 355:1858– 1863. Nicolich MJ, Huebner WW, Schnatter AR. Influence of parental and biological factors on the male birth fraction in the United States: an analysis of birth certificate data from 1964 through 1988. Fertil Steril 2000;73:487 – 492. Orvos H, Kozinszky Z, Bartfai G. Natural variation in the human sex ratio. Hum Reprod 2001;16:803. Parazzini F, Chatenoud L, Maffioletti C, Chiaffarino F, Caserta D. Periconceptional smoking and male:female ratio of newborns. Eur J Public Health 2005;15:613 – 614. Peippo J, Bredbacka P. Sex-related growth rate differences in mouse preimplantation embryos in vivo and in vitro. Mol Reprod Dev 1995; 40:56– 61. Rubessa M, Boccia L, Campanile G, Longobardi V, Albarella S, Tateo A, Zicarelli L, Gasparrini B. Effect of energy source during culture on in vitro embryo development, resistance to cryopreservation and sex ratio. Theriogenology 2011;76:1347 – 1355. Rueness J, Vatten L, Eskild A. The human sex ratio: effects of maternal age. Hum Reprod 2012;27:283 – 287. Ryan JJ, Amirova Z, Carrier G. Sex ratios of children of Russian pesticide producers exposed to dioxin. Environ Health Perspect 2002;110:A699–A701. Setti AS, Figueira RC, Braga DP, Iaconelli A Jr, Borges E Jr. Gender incidence of intracytoplasmic morphologically selected sperm injection-derived embryos: a prospective randomized study. Reprod Biomed Online 2012;24:420–423. Tarin JJ, Garcia-Perez MA, Hermenegildo C, Cano A. Changes in sex ratio from fertilization to birth in assisted-reproductive-treatment cycles. Reprod Biol Endocrinol 2014;12:56. Tiffin GJ, Rieger D, Betteridge KJ, Yadav BR, King WA. Glucose and glutamine metabolism in pre-attachment cattle embryos in relation to sex and stage of development. J Reprod Fertil 1991;93:125 – 132. Torner E, Bussalleu E, Briz MD, Yeste M, Bonet S. Embryo development and sex ratio of in vitro-produced porcine embryos are affected by the energy substrate and hyaluronic acid added to the culture medium. Reprod Fertil Dev 2014;26:570 – 577. Vatten LJ, Skjaerven R. Offspring sex and pregnancy outcome by length of gestation. Early Hum Dev 2004;76:47 – 54. Weston G, Osianlis T, Catt J, Vollenhoven B. Blastocyst transfer does not cause a sex-ratio imbalance. Fertil Steril 2009;92:1302 – 1305. Zhu J, Li M, Chen L, Liu P, Qiao J. The protein source in embryo culture media influences birthweight: a comparative study between G1 v5 and G1-PLUS v5. Hum Reprod 2014;29:1387– 1392.

Downloaded from http://humrep.oxfordjournals.org/ at University of Aberdeen on April 15, 2015

Catalano RA. Sex ratios in the two Germanies: a test of the economic stress hypothesis. Hum Reprod 2003;18:1972– 1975. Dean JH, Chapman MG, Sullivan EA. The effect on human sex ratio at birth by assisted reproductive technology (ART) procedures—an assessment of babies born following single embryo transfers, Australia and New Zealand, 2002 – 2006. BJOG 2010;117:1628– 1634. Dyrlund TF, Kirkegaard K, Poulsen ET, Sanggaard KW, Hindkjaer JJ, Kjems J, Enghild JJ, Ingerslev HJ. Unconditioned commercial embryo culture media contain a large variety of non-declared proteins: a comprehensive proteomics analysis. Hum Reprod 2014;29:2421 – 2430. Fukuda M, Fukuda K, Shimizu T, Moller H. Decline in sex ratio at birth after Kobe earthquake. Hum Reprod 1998;13:2321– 2322. Fukuda M, Fukuda K, Shimizu T, Andersen CY, Byskov AG. Parental periconceptional smoking and male: female ratio of newborn infants. Lancet 2002;359:1407– 1408. Griffin DK, Handyside AH, Harper JC, Wilton LJ, Atkinson G, Soussis I, Wells D, Kontogianni E, Tarin J, Geber S et al. Clinical experience with preimplantation diagnosis of sex by dual fluorescent in situ hybridization. J Assist Reprod Genet 1994;11:132– 143. Gutierrez-Adan A, Oter M, Martinez-Madrid B, Pintado B, De La Fuente J. Differential expression of two genes located on the X chromosome between male and female in vitro-produced bovine embryos at the blastocyst stage. Mol Reprod Dev 2000;55:146 – 151. Gutierrez-Adan A, Granados J, Pintado B, De La Fuente J. Influence of glucose on the sex ratio of bovine IVM/IVF embryos cultured in vitro. Reprod Fertil Dev 2001;13:361 – 365. Hesketh T, Xing ZW. Abnormal sex ratios in human populations: causes and consequences. Proc Natl Acad Sci USA 2006;103:13271– 13275. Jacobsen R, Moller H, Mouritsen A. Natural variation in the human sex ratio. Hum Reprod 1999;14:3120 – 3125. Kimura K, Spate LD, Green MP, Roberts RM. Effects of D-glucose concentration, D-fructose, and inhibitors of enzymes of the pentose phosphate pathway on the development and sex ratio of bovine blastocysts. Mol Reprod Dev 2005;72:201– 207. Kimura K, Iwata H, Thompson JG. The effect of glucosamine concentration on the development and sex ratio of bovine embryos. Anim Reprod Sci 2008;103:228 – 238. Lin PY, Huang FJ, Kung FT, Wang LJ, Chang SY, Lan KC. Comparison of the offspring sex ratio between cleavage stage embryo transfer and blastocyst transfer. Taiwan J Obstet Gynecol 2010;49:35 – 39. Lin S, Li M, Lian Y, Chen L, Liu P. No effect of embryo culture media on birthweight and length of newborns. Hum Reprod 2013;28:1762 – 1767. Luke B, Brown MB, Grainger DA, Baker VL, Ginsburg E, Stern JE. The sex ratio of singleton offspring in assisted-conception pregnancies. Fertil Steril 2009; 92:1579– 1585. Luna M, Duke M, Copperman A, Grunfeld L, Sandler B, Barritt J. Blastocyst embryo transfer is associated with a sex-ratio imbalance in favor of male offspring. Fertil Steril 2007;87:519 – 523.

7