Effects of in Vitro Fertilization (IVF) Therapies on Metabolic, Endocrine and Inflammatory Status in IVF-Conceived Pregnancy

Ayla Coussa; Hayder A. Hasan; Thomas M. Barber


Clin Endocrinol. 2020;93(6):705-712. 

In This Article



A total of 702 women were prescreened by telephone, from whom 673 were eligible for enrolment in the study. Of these, 354 women were recruited into the study as they had ET. Post–embryo transfer, biochemical and ectopic pregnancies (n = 10) were excluded and participants were divided into two groups: clinical successful pregnancy based on positive β-HCG (pregnant, n = 191) and unsuccessful pregnancy outcome based on negative serum β-HCG (nonpregnant, n = 153). In the pregnant group (n = 191), 16 participants did not complete the study, 17 experienced a miscarriage (8%) before 12 weeks and they were hence excluded from the study. In total, 275 participants completed the study, of which the two groups included successful IVF-conceived pregnancy (n = 158) and unsuccessful IVF pregnancy (n = 117; 36 participants did not complete the study). Successful pregnancy accounted for 105 single (66%) and 53 multiple pregnancies (34%). An overview of the recruitment flow is shown in Figure 2. The reasons for undergoing IVF therapy were as follows: 30% female infertility, 30% male infertility, 15% for gender selection and 25% related to other causes (genetic disorders, combination of both male and female infertility and unknown infertility). Overall, 45% of women undergoing IVF had a confirmed diagnosis of polycystic ovary syndrome (PCOS) (n = 122; 50 nonpregnant and 72 pregnant), but this was not necessarily the main reason for undergoing IVF in these cases. For the pregnant group overall at baseline, 22% of women had a BMI ≥ 30 kg/m2 (n = 22), 16% had HbA1c in the prediabetic range between 5.7–6.1% (n = 26), and 14% had a past history of GDM (n = 22).

Figure 2.

Flow chart of participants' recruitment and enrolment in the study

At baseline, there was no difference in anthropometrics, endocrine and metabolic parameters between pregnant and nonpregnant women (data shown in Table 1). Participants overall had a median age of 32 (6) years, BMI of 25.4 (6.9) kg/m2, HbA1c of 5.2 (0.52) % and TSH of 1.82 (1.4) μIU/mL. At baseline, 21% had a BMI ≥ 30 kg/m2. Ethnicity of participants was multicultural: 53% Gulf nationals, 20% from Far East (South and East Asia), 15% Middle Eastern, 8% Europeans and 4% with African origins. There was a significant increment in weight at 12 weeks in both groups: pregnant women BMI: 24.8–25.7 kg/m2, P < .001; nonpregnant women BMI: 25.6–25.8 kg/m2, P = .002.

Glucose and Insulin Homeostasis

Compared with baseline, glucose level was significantly lower at 4 weeks (Δ: −1.68 mg/dL, P = .007) and more so at 12 weeks (Δ: −3.96 mg/dL, P = .001) in the pregnant group. At week 12, pregnant women had also significant reduction in HbA1c (5.3 to 5.08%, P < .001). There were no changes in fasting insulin and HOMA-IR measures at 4 and 12 weeks of pregnancy compared with baseline levels. In contrast, nonpregnant 4-week glucose level did not change compared with baseline, while insulin (Δ: 3.1 μIU/mL, P < .001) and HOMA-IR measures (Δ: 0.7, P = .01) were increased. At 12 weeks, nonpregnant women showed statistically significant increase in glucose (Δ: 1.56 mg/dL), while insulin (Δ: −1.17 μIU/mL) and HOMA-IR (Δ: −0.2) were slightly reduced at week 4 but remained higher compared to baseline levels, with P < .001 (Table 1 and Figure 3). Regardless of pregnancy status, significant changes in glucose and insulin homeostasis occur at 4 weeks of IVF therapy and dependent on IVF outcome. In relation to history of PCOS, in the pregnant group, 72 women had a history of PCOS and 86 were non-PCOS prior to pregnancy. At baseline, women with PCOS presented higher insulin (11.3 vs 9.5 μIU/mL) and HOMA-IR (2.5 vs 2.05) levels compared to the non-PCOS group, with P < .001. At 12 weeks of pregnancy, there were no differences in insulin homeostasis parameters between the two groups of pregnant women. In the nonpregnant group, baseline and 12-week insulin homeostasis did not differ between PCOS and non-PCOS women.

Figure 3.

Comparison of glucose homeostasis at baseline and 12 wk of IVF therapy between pregnant and nonpregnant women (Graph A. Fasting glucose level; B. Fasting insulin level; C. HOMA-IR). *P < .05 vs at 12 wk of IVF therapy; Ŧ P < .05 vs pregnancy; HOMA-IR, homeostatic model assessment of insulin resistance; IVF, in vitro fertilization


At 12 weeks of pregnancy, lipid profile increased significantly, including T-Chol (177.5–199.5 mg/dL), TG (73.5–126.8 mg/dL) and HDL-C (55.3–65.1 mg/dL), with P < .001. Similarly, in nonpregnant women, T-Chol (169.5–174.9 mg/dL), TG (71.0–83.7 mg/dL) and HDL-C (52.0–54.1 mg/dL) increased (P < .001).

Endocrine and Inflammatory Outcomes

Pregnant women had significant reductions in TSH (1.7–1.4 μIU/mL; P < .001) at 12 weeks, while for nonpregnant group, there was no change in TSH level. For all subjects overall, and for each pregnant and nonpregnant groups, there were no changes in serum adiponectin or LBP between baseline and 12-weeks.