Obesity and adverse pregnancy outcomes in older patients with decreased ovarian reserve: a retrospective single-centre study

Background: In recent years, infertility has increased in older women with 15 decreased ovarian reserve (DOR). Studies have shown that women with DOR have 16 fewer oocytes, which are poorer in quality, and have an increased risk of adverse 17 pregnancy outcomes. Pre-pregnancy BMI is significantly correlated with many 18 adverse pregnancy outcomes. Therefore, we conducted this study to explore the 19 correlation between body mass index (BMI) and abortion and live birth in older 20 patients with DOR. 21 Methods : The clinical data of 2052 older women with infertility and DOR admitted to the Reproductive Medicine Center of the First Affiliated Hospital of 23 Zhengzhou University from August 2009 to May 2018 were analysed retrospectively. 24 Patients were divided into underweight (BMI < 18.5 kg/m 2 ; n = 56), normal weight 25 (18.5 kg/m 2 ≤ BMI < 24 kg/m 2 ; n = 1389), overweight (24 kg/m 2 ≤ BMI < 28 kg/m 2 ; n = 26 527) and obese (BMI ≥ 28 kg/m 2 ; n = 80). We compared the pregnancy outcomes 27 of patients in each group. 28 Results: Logistic regression analysis showed that being overweight or obese 29 were independent risk factors for miscarriage (P < 0.05) and protection factors for live 30 births (P ＜ 0.05). Being underweight was a protective factor for live births (P < 0.05). 31 Conclusions: The abortion and live birth rates in older infertile women with 32 DOR are correlated with BMI. Higher BMI was associated with higher abortion rates 33 and lower live birth rates. Being underweight also correlated with the live birth rate. Therefore, to improve pregnancy outcomes, we suggest that older patients with DOR 35 may benefit from maintaining a normal weight before seeking fertility treatments.


Conclusions:
The abortion and live birth rates in older infertile women with 32 DOR are correlated with BMI. Higher BMI was associated with higher abortion rates 33 and lower live birth rates. Being underweight also correlated with the live birth rate. 34 Therefore, to improve pregnancy outcomes, we suggest that older patients with DOR 35 may benefit from maintaining a normal weight before seeking fertility treatments. Ovarian reserve is the capacity for growth and development of follicles in the 42 female ovarian cortex and the ability to form fertilised oocytes. Diminished ovarian 43 reserve (DOR) is a common endocrine disease in women of childbearing age and 44 refers to the decline in the number and quality of oocytes, ovulation disorders, 45 endocrine disorders, and infertility due to factors such as age, metabolism, genetics, 46 autoimmunity, iatrogenicity, toxicity, and infection. In the process of assisted 47 reproductive technology (ART), DOR is characterised by poor drug response, few 48 eggs, low number of high-quality embryos, high rate of cycle cancellation, and low 49 clinical pregnancy rate [1]. 50 Studies have shown that pre-pregnancy BMI is significantly correlated with many 51 adverse pregnancy outcomes, such as gestational diabetes mellitus (GDM), 52 hypertensive disorders in pregnancy, premature birth, abnormal birth weight, and 53 cesarean section [2,3]. However, there are currently insufficient data on the role of 54 BMI in pregnancy outcomes in patients with decreased ovarian reserve. Therefore, 55 we conducted this study to explore the relationship between BMI and pregnancy 56 outcomes in patients with DOR to provide a reference for clinical practice.  3) Specimen collection and laboratory tests: In the patient's natural physiological 89 state, the second to fourth days of the menstrual cycle or menopause for more than 90 50 days (excluding early pregnancy and B-ultrasound monitoring of the ovaries and 91 endometrium are consistent with anovulatory status)，3 ml of venous blood was 92 drawn on an empty stomach, serum was collected by centrifugation, and 93 electrochemiluminescence immunoassay kit (Roche, Germany) was used to detect 94 serum basal luteinising hormone (bLH), basal follicle stimulating hormone (bFSH), 95 and anti-Mullerian hormone (AMH) levels (inter-and intra-batch detection difference: 96 < 5%). 97 4) ART protocol: a gonadotropin (Gn) releasing hormone (Gn) agonist was used 98 to prevent a premature surge in luteinising hormone (LH), and Gn was used to 99 stimulate follicular growth. When the largest follicle diameter was greater than 20 100 mm, and more than 2/3 of the total follicles were >16 mm. Human chorionic 101 gonadotropin (hCG) was administered according to the serum FSH, LH, E2 and P 102 levels. Ultrasound-guided egg retrieval was performed 36-38 hours later. 103 5) Outcome indicators: At 14 or 18 days after embryo transfer, serum β-hCG 104 levels were measured to detect early pregnancy. Ultrasonography was performed 35 105 or 45 days after embryo transfer, and we diagnosed pregnancy clinically by the 106 existence of an intrauterine pregnancy sac and a positive heartbeat. Miscarriage was 107 defined as termination of pregnancy before 28 weeks' gestation with a foetal weight 108 of less than 1000 g. Live birth was defined as at least one live birth after 24 weeks of 109 pregnancy. We defined other outcomes as follows: implantation rate = number of 110 gestational sacs / number of embryos transferred × 100%; clinical pregnancy rate = 111 number of clinical pregnancy cycles / total number of transplanted cycles × 100%; 112 abortion rate = number of abortion cycles / total number of pregnancy cycles × 100%; 113 and live birth rate = number of live birth cycles / total number of transplant cycles × 114 100%. 115 3. Statistical analysis was performed using SPSS 22.0 (IBM Corp., Armonk, NY, 116 USA) statistical software for data analysis. Normally distributed data are expressed 117 as mean ± standard deviation (x±s), one-way ANOVA was used for comparison 118 between groups. Continuous variables with skewed distributions are represented as 119 medians (interquartile ranges, IQR), and were compared using the Kruskal-Wallis 120 test. Count data were expressed as rate (%), and the chi-square test was used to 121 compare groups (X2). The difference of proportions between groups was compared 122 using Bonferroni correction. Binary logistics regression was used to determine the 123 correlation between BMI and pregnancy outcomes (abortion and live birth rates). The 124 results are presented as the adjusted odds ratios (aORs) with the 95% confidence 125 intervals (CIs). Statistical significance was set at P < 0.05.

Baseline data 132
There were significant differences in male age, female age, menstrual cycle 133 length, bFSH levels, bLH levels, AMH levels, and AFC among the different BMI 134 classifications (all P < 0.05). Menstrual cycle length was directly proportionate to 135 increased BMI. There were no significant differences in male BMI level, infertility 136 diagnosis, and previous in vitro fertilization/intracytoplasmic sperm injection (IVF/ICSI) 137 attempts (all P > 0.05) ( Table 1). 138

Analysis of patients' transplant status and assisted pregnancy results 139
The relationship between BMI and transplant and fertility outcomes of patients with 140 reduced ovarian reserve was analysed. There was no significant correlation between 141 BMI and initial Gn dose, Gn dosage, endometrial thickness on the day of HCG 142 administration, number of retrieved oocytes, number of available embryos, number of 143 embryos transferred, and embryo stage at transfer (all P > 0.05). However, we did find 144 a significant correlation between BMI and the abortion rate (P = 0.015) and live birth 145 rate (P = 0.016). There was no significant correlation between BMI and implantation 146 rate, clinical pregnancy rate, number of live births, Cumulative clinical pregnancy rate 147 (CCPR) or Cumulative live birth rate (CLBR) (all P > 0.05) ( Table 2). 148

Analysis of factors affecting miscarriage 149
Using binary logistic regression to analyse related factors, we found that before 150 adjusting for confounding factors, male age, female age, and being overweight were 151 independent risk factors for miscarriage. After adjusting for male age, female age, 152 menstrual cycle, bFSH, bLH, AMH, and AFC, only being overweight (adjusted odds 153 ratio [aOR] = 2.41; 95% confidence interval [CI]: 1.20-4.83; P = 0.013) or obese (aOR 154 = 6.41; 95% CI: 1.38-29.70; P = 0.018) was independently associated with 155 miscarriage, with the aOR value of the obesity group found to be several times that of 156 the overweight group (Table 3; Figure 2A). 157

Analysis of factors correlated with abortion and live birth 158
Using binary logistic regression analysis, we found that male age, female age, 159 and being overweight were independently associated with abortion before adjusting 160 for confounding factors. After adjusting for male age, female age, menstrual cycle 161 length, BMI, bFSH, bLH, AMH, and AFC, we found that factors such as being

Lifestyle changes and pregnancy outcomes 233
Maternal obesity increases the risk of pregnancy complications such as GDM, 234 gestational hypertension, and preeclampsia [31]. In addition, more than half of 235 overweight and obese women gain more weight than recommended during 236 pregnancy, which leads to an increased risk of perinatal complications and poor 237 neonatal outcomes, and affects the health of the mother and future generations [32]. 238 Studies have shown that female obesity is an independent risk factor in the 239 cumulative live birth rate in the first complete ovarian stimulation cycle [ The present study findings suggest that female obesity is an independent risk 255 factor for abortion in older patients with DOR, with greater risk in obese women than 256 in overweight women. In women with normal weight, BMI is an independent 257 protective factor in the live birth rate. Considering the difficulty experienced by 258 women with DOR in conceiving and remaining pregnant, and the high obesity rate in 259 older women, we recommend that women reduce their pre-pregnancy weight through 260 lifestyle changes. 261

Advantages and limitations 262
Our study presents a novel correlation of pregnancy outcomes in IVF/ICSI-ET 263 with BMI. We have attempted to control for confounding factors that affect pregnancy 264 outcomes as much as possible to improve the reliability of our results. Although we 265 have reduced selection and confounding biases as much as possible, the present 266 study is a retrospective study with inherent limitations. Our sample size for the 267 underweight and obese patients is small. The study should be repeated with a larger 268 sample size. In addition, this study is a single-centre study, and we only used the 269 clinical data from recent transplant cycles of all older patients DOR in the same 270 centre. Our study lacks some advantages of multi-centre research; however, single-271 centre research can arguably provide more consistent results by avoiding 272 inconsistencies in surgical methods and laboratory conditions. Finally, we did not 273 evaluate cumulative pregnancy outcomes or neonatal and obstetric outcomes, which 274 may present opportunities for future research. For infertile women > 35 years old with reduced ovarian reserve, pregnancy 278 outcomes of IVF/ICSI-ET were correlated with BMI. We found that BMI above the 279 normal range was correlated with an increased risk of miscarriage. Being 280 underweight or overweight was also associated with the live birth rate. Obesity was 281 more strongly associated with abortion and reduced live birth rate than being 282 overweight. Our findings suggest that older patients with DOR who wish to conceive 283 may benefit from maintaining a normal BMI to improve pregnancy outcomes during 284 fertility treatment.  Hospital of Zhengzhou University, and all patients signed an informed consent form. 312 All methods were conducted in accordance with relevant guidelines and regulations. 313

Consent for publication 314
Not applicable. 315

Availability of data and materials 316
The datasets used in the current study are available from the corresponding 317 author on reasonable request. 318

Competing interests 319
The authors declare that they have no competing interests 320