Impact of Central Obesity on Women with Polycystic Ovary Syndrome Undergoing In Vitro Fertilization

Abstract Central obesity (CO) is a defining characteristic of polycystic ovary syndrome (PCOS) and PCOS-induced disorders are likely to be exacerbated in the presence of CO. This study aims to evaluate the impact of CO on infertile women with PCOS undergoing in vitro fertilization (IVF).It is a retrospective and case–control study. One hundred eighty-eight infertile PCOS women undergoing IVF were divided into CO group (n = 70, waist circumference [WC] ≥80 cm) and noncentral obesity (NCO) group (n = 118, WC <80 cm). Baseline characteristics, parameters of ovarian stimulation and laboratory, and pregnancy outcomes were compared between two groups. After controlling for body mass index (BMI), WC positively correlated with fasting insulin (r = 0.210, p = 0.007), homeostatic model assessment for insulin resistance (r = 0.249, p = 0.006) and free androgen index (r = 0.249, p = 0.006). Compared with NCO group, CO group had significantly increased endocrine and metabolic disorders and needed significantly higher dose of gonadotropins, longer duration of ovarian stimulation (p < 0.05), but had significantly lower peak serum estradiol level (p < 0.01) and less oocytes retrieved (p = 0.032). CO group had significantly lower live birth and implantation rates (53.8% vs. 86.8%, p = 0.001; and 24.3% vs. 36.3%, p = 0.019, respectively) and higher early spontaneous miscarriage rate (38.5% vs. 7.5%, p = 0.002). For the multivariate analysis, by adjusting for age, BMI, insulin resistance, and hyperandrogenism (HA), CO was significantly independent risk factor for early miscarriage (adjusted relative ratio = 16.87, 95% confidence interval = 2.15–132.70, p = 0.007). CO is associated with insulin resistance, hyperinsulinemia, and HA independent of BMI and is associated with poor pregnancy outcome in infertile women with PCOS undergoing IVF.


Introduction
Polycystic ovary syndrome (PCOS) is the most common and complicated endocrine disorder in women of reproductive age, with multiple endocrine and metabolic disorders. Subfertile women with PCOS will usually benefit from conventional treatments, such as lifestyle changes, laparoscopic ovarian drilling, or ovulation in-duction. 1 However, when undergoing in vitro fertilization (IVF)/intracytoplasmic sperm injection (ICSI), PCOS women may suffer from high rates of cycle cancellations, spontaneous miscarriage, and decreased rate of fertilization. [2][3][4] PCOS women represent a cohort of people with a higher prevalence of overweight and obesity compared with healthy women, 5 although there is a wide variability in the estimates of obesity in PCOS women across different countries and ethnicities. 6 The prevalence ranges from 66% to 80% in PCOS women from the Western populations 7,8 and about 23% in Chinese PCOS women. 9 It is found that overweight and obesity are closely related to the severity of endocrine disorders, such as insulin resistance, impaired glucose tolerance, and other metabolic abnormalities in PCOS women. [10][11][12] Furthermore, it is also important to emphasize that visceral, rather than subcutaneous, fat, also named central obesity (CO), is a defining characteristic of PCOS. 13 CO can induce local and systemic oxidative stress in PCOS patients, which suggest that PCOS-induced disorders are likely to be exacerbated in the presence of CO. 14 The present study aims to evaluate whether CO is associated with ovarian stimulation and pregnancy outcome in Chinese infertile PCOS women undergoing IVF/ICSI.

Study population
This was a retrospective study. From January 2009 to December 2010, PCOS women who accepted their first fresh IVF/ICSI cycles at the IVF center of Sun Yat-sen Memorial Hospital, Sun Yat-sen University were included. This study was approved by the local Medical Ethics Committee.
The PCOS patients were diagnosed according to the 2003 Rotterdam diagnostic criteria. 15 Patients were excluded if they had uterine malformation, untreated intrauterine lesions, and previous ovarian surgery. Women with waist circumference (WC) more than 80 cm were defined as CO based on the 2005 International Diabetes Foundation consensus (http://idf.org/home). Patients were divided into CO group and noncentral obesity (NCO) group.
The phenotypes in PCOS women were divided according to previous studies. 16,17 Following these studies, PCOS features were defined hyperandrogenism (HA), oligoanovulatory ovarian dysfunction (OAD), and polycystic ovarian morphology (PCOM). Then four different PCOS phenotypes were identified: phenotype A or complete phenotype (HA, OAD, and PCOM), phenotype B or non-PCOM phenotype (HA and OAD without PCOM), phenotype C or ovulatory PCOS phenotype (HA and PCOM with ovulatory cycles), phenotype D or normoandrogenic PCOS (OAD and PCOM without HA).

Clinical and biochemical measurements
Data on characteristic features, ultrasonographic, and laboratory variables were collected. The height, weight, WC, and hip circumference of a patient were measured according to World Health Organization recommendation. 18 Transvaginal ultrasonography was performed on day 3-5 of menstrual cycle or after a progestininduced withdrawal bleeding to exclude any pelvic pathology and to determine the antral follicle count (AFC  19 The homeostatic model assessment for insulin resistance (HOMA-IR) and free androgen index (FAI) were calculated as following: (fasting plasma insulin [mIU/L] · fasting plasma glucose [mmol/l])/22.5, and (100 · TT [nmol/L])/SHBG [nmol/L]), respectively. Insulin resistance was defined as HOMA-IR >2.14 according to our previous study. 20 IVF/ICSI and embryo transfer procedures All PCOS patients were pretreated according to the consensus on PCOS patients' treatment. 21 When patients were defined as insulin resistant, IFG, or IGT, metformin was administrated for 1-3 month. Standard controlled ovarian stimulation protocol was performed as follows. 22 All patients were treated with oral contraceptive pills (Yasmin, Scherring, Germany) from cycle days 3 or days 3 after a progestin-induced withdrawal bleeding.
Then long-acting triptorelin acetate (Diphereline, Ipsen, France) was given at the 17th-19th day with a single dose of 1.25 mg subcutaneously. Fourteen days after triptorelin acetate was given, when the endometrium with the thickness of £5 mm and suppressed ovaries (no antral follicles ‡10 mm) were detected by ultrasound scan and serum estradiol levels £50 pg/mL and LH levels £5 IU/L were confirmed, ovarian stimulation was started. Around 112.5-225 IU/day recombinant FSH (Gonal-F; Merck Serono, Germany) was administered according to the patient's age, AFC, and basal FSH level. If necessary, the dosage of recombinant FSH was adjusted or 75-150 IU/day recombinant LH (Luveris; Merck Serono, Germany) was added according to ovarian response.
Urinary human chorionic gonadotropin (hCG; Lizhu, China) was given for triggering when at least two follicles had reached a diameter of ‡18 mm or three follicles had reached a diameter of ‡17 mm. Around 4000-10,000 IU of hCG was administrated depending on follicular numbers, peak E 2 level, and body mass index (BMI). Transvaginal oocyte retrieval was scheduled 36-37 h after the hCG injection.
Then, standard laboratory procedures for conventional IVF/ICSI were followed. 23 Evaluation of the quality of embryos was performed on day 3 according to ''embryo grading''. 24 Embryos of modified Hu's grades 1 or 2 were considered of high quality (7-9 cell; blastomeres of equal size; 0-20% cytoplasmic fragments). When available, two embryos were transferred in young women (<35 years) and three embryos were transferred in older women ( ‡35 years). Embryo transfers (ET) were performed under ultrasound guidance with a full bladder using a Wallace catheter (Wallace Ltd, Colchester, England). Luteal phase support was sustained with natural progesterone in oil 60 mg intramuscular injection daily from the day of oocyte retrieval.

Pregnancy outcomes
Biochemical pregnancy was defined as elevated serum b-hCG ( ‡25 IU/L) 14 days after ET. Clinical pregnancy was defined as the presence of gestational sac(s) by ultrasonography 4-5 weeks after ET.

Statistical analysis
Statistical analysis of the data was performed using SPSS version 21.0 (SPSS, Inc., Chicago, IL). Continuous variables were compared using Student's t-test or analysis of variance for parametric data and Mann-Whitney test or Kruskal-Wallis test for nonparametric data. Categorical data were compared using Chisquared test. Pearson partial correlation was used for correlation analysis between two continuous variables after adjusting some confounding factor.

Baseline characteristics
A total of 188 women with PCOS met the eligibility criteria. There were 70 and 118 women within the CO and NCO groups, respectively.
The baseline characteristics of the subjects are shown in Table 1. When comparing CO with NCO group, BMI, WC, hip circumference, waist-hip ratio, FAI, fasting glucose, fasting insulin (FIN), HOMA-IR, 2-h glucose and insulin in oral glucose tolerance test, serum TG, and the occurrence rates of IR, IFG, IGT were significantly higher, whereas serum SHBG and HDL-C was significantly lower ( p < 0.05). There was no significant difference in age, duration of infertility, baseline serum hormones (FSH, LH, TT, FT, DHEA-S), Serum TC, LDL-C, bilateral volumes of ovaries, AFC, and rates of primary infertility and DM2 between the two groups ( p > 0.05). In addition, the PCOS phenotypes were not significant between the two groups ( p > 0.05).

Parameters during IVF/ICSI
The detailed parameters during IVF/ICSI treatment are shown in Table 2. One cycle was cancelled in CO group, whereas three in NCO group due to poor ovar-ian response. Compared with NCO group, CO group needed significantly higher dose of gonadotropin, longer duration of ovarian stimulation, higher dose of hCG for trigger( p < 0.05), but had significantly lower peak serum estradiol level, less mature follicles on trigger day, and less oocytes retrieved ( p < 0.05). There were no significant differences in endometrial thickness, cleavage rate, number of frozen embryos, percentage of embryos available, number of high-quality embryos, and rates of completed cycles, ICSI cycles, and fertilization rate between the two groups.
Multivariate analysis for early miscarriage showed that when compared with NCO, the crude RR for CO was 7.50 (95% CI = 2.06-27.25, p = 0.002) and remained statistically significant (adjusted RR = 16.87, 95% CI = 2.15-132.70, p = 0.007) after adjusting for age, BMI, insulin resistance, and HA (Table 4). In addition, the RR for age Values are expressed as mean -SD or numbers (percentages). *Significant differences between two groups, p < 0.01. **Significant differences between two groups, p < 0.05. E 2 , estradiol; hCG, human chorionic gonadotropin; ICSI, intracytoplasmic sperm injection. Values are expressed as mean -SD or numbers (percentages). *Significant differences between two groups, p < 0.05. **Significant differences between two groups, p < 0.01. Biochemical pregnancy rate: biochemical pregnancy cycles/embryo transfer cycles; clinical pregnancy rate: clinical pregnancy cycles/embryo transfer cycles; implantation rate: no. of implantation gestational sac/no. of total embryos transferred; multiple pregnancy rates: multiple pregnancy cycles/clinical pregnancy cycles; early miscarriage rate: early miscarriage cycles/clinical pregnancy cycles; ectopic pregnancy rate: ectopic pregnancy cycles/clinical pregnancy cycles; live birth rate: live birth cycles/clinical pregnancy cycles.

Discussion
PCOS is the most common and complicated endocrine dysfunction in women of childbearing age. A majority of PCOS patients are obese or have a normal BMI, but are metabolically obese with WC more than 80 cm. Pregnancy achievement and maintenance is adversely affected by obesity, overweight, or elevated BMI. Huang et al. reported that obese PCOS patients obtained lower clinical pregnancy and live birth rates. 25 However, WC may be a good indicator to identify insulin resistance and metabolic syndrome, especially among population of BMI £30 (kg/m 2 ). 26,27 WCs seem to help identify those at increased health risk within the normal weight, overweight, and class I obese BMI categories. 26 This study was to evaluate the effects of CO on the outcomes of IVF/ICSI in women with PCOS. Among these patients in this study, 98.9% women had BMI £30 kg/m 2 ; CO was associated with endocrinal and metabolic disorders in women with PCOS and was significantly associated with insulin resistance, hyperinsulinemia, and HA independent of BMI. This suggested that WC could be a better marker than BMI for Chinese women, which is in accordance with a study conducted by Stepto et al. 28 Therefore, WC could be a good and simple indicator for distinguishing patients with different metabolic status in PCOS women, to CO and NCO groups.
Previous studies had shown the PCOS phenotypes played an important role in the variability of pregnancy complication 16 and oocyte competence. 17 Therefore, in this study, patients were subclassified according to the four phenotypes, and there were no significant differences between the two groups.
Obese PCOS patients needed more dose of gonadotropins, had less mature follicles on hCG day, and less oocytes retrieved after COS. [29][30][31] Altered pharmacokinetics of gonadotropin and high intrafollicular leptin concentration may be related to gonadotropin resistance in obese PCOS women. 30 Patients with gonadotropin resistance might have increased the rate of cycle cancellations due to poor ovary responses in common population. However, the condition of increased cycle cancellations did not appear in our CO group. It may be because more AFC was found in PCOS patients, which can recruit more follicles after adjusting the dose of gonadotropin during ovarian stimulation. CO patients accepted higher dose of gonadotropin, which may increase follicle recruitment and oocytes retrieved.
The present study indicated that PCOS women in CO group had decreased clinical pregnancy rate, significantly higher early spontaneous miscarriage rate, and lower live birth rate compared with the NCO group. Insulin resistance may be responsible for this condition. 32,33 Moreover, in CO group, the number of highquality embryos, and rate of biochemical pregnancy were lower than NCO group, although it was not significant. In multivariate analysis we could see that CO was a significantly independent risk factor for early miscarriage. It remained significant after adjusting for insulin resistance, which indicated that PCOS infertile women with CO in China were at a greater risk in early spontaneous miscarriage.
Possible mechanism of poor pregnancy outcome in PCOS women with CO is that metabolic disturbance such as insulin resistance result in abnormalities during folliculogenesis, follicular growth, oocyte meiotic maturation processes, and uterine receptivity, through circulating endocrine and local paracrine/autocrine mechanisms. [34][35][36] With regard to incidence of moderate OHSS and freeze-all cycles due to the risk of OHSS, the NCO group showed higher rates. It can be explained that lean PCOS patients are a high-risk group of OHSS in itself. 37 The special strength of this study was that CO could be a better indicator to identify the comprehensive endocrinal and metabolic disorders and poor pregnancy outcome undergoing IVF treatment among Chinese women with PCOS, who have much higher proportion of normal weight, overweight, and class I obese BMI categories than western population.
The main limitation of this study is the relatively small sample size and retrospective nature. Further studies in larger populations are needed to come up with more supporting data. Furthermore, because of differences in ethnicity, effects of the CO on PCOS women could account for some conflicting results. However, to our knowledge, this study is the first to identify the association of CO and pregnancy outcome in PCOS women undergoing IVF/ICSI.
In summary, PCOS patients with CO have severe insulin resistance and metabolic disorders independent of BMI. When undergoing IVF, they need more dose of gonadotrophin and longer duration of ovarian stimulation, but have lower number of retrieved oocytes. PCOS patients with CO have higher early miscarriage rate, reduced implantation, and live birth rates after IVF/ICSI.

Details of Ethics Approval
This study was approved by the Ethics Committee of Sun Yat-sen Memorial Hospital, Sun Yat-sen University (No. 201408).