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


This study showed that IVF hormonal therapy raises levels of glucose, insulin and lipids parameters and reduces insulin sensitivity in women with failed IVF (ie, nonpregnant). A similar but more intense effect on impaired insulin sensitivity and glucose intolerance was reported with long-term use of oral contraceptives,[24,25] suggesting that the duration of the treatment might have more impact on glucose and insulin homeostasis than a high dose for a short period of time. The effect of IVF hormones on glucose and insulin homeostasis was down-regulated by pregnancy, with no change observed in insulin level and resistance (HOMA-IR), while glucose level was reduced (still within normal range). Glucose level may drop early in pregnancy, secondary to a physiological adaptation for increased foetoplacental needs, with a focus on carbohydrates as a source of energy.[6,26] During early gestation, glucose homeostasis remained similar to nonpregnancy level, confirming also other studies' findings conducted in spontaneous pregnancy in maintaining euglycaemic levels.[27,28] The diabetogenic state from hyperinsulinaemia and increased insulin resistance is usually manifested during mid-preganncy;[27] such changes were not observed yet at 12 weeks of IVF-conceived pregnancy. In relation to history of PCOS, despite higher baseline levels of insulin resistance markers (insulin and HOMA-IR; levels still within normal range) in PCOS women, 12-week levels were down-regulated to non-PCOS pregnant levels.

Unlike the expected effect of oestrogen therapy on raising TSH, level remained the same at week 12 for nonpregnant women. Two possible explanations may be proposed: (a) IVF medications were stopped at 4 weeks and oestrogen therapy has already been cleared out from the body; (b) duration of IVF hormone administration was too short to induce changes in TSH level. Drop in TSH level in the pregnant group is consistent with previous studies reporting 20–50% suppression due to the sharp increase in hCG concentrations.[29,30] Given the potential adverse effect of IVF hormones on impairing glucose and insulin homeostasis and thyroid function, more attention should be paid with repeated IVF cycles or if IVF hormones were to be provided for a longer period during pregnancy.

Female reproductive hormones stimulate synthesis of inflammatory markers, which may be associated with change in gut permeability;[13] a similar response was expected with IVF hormonal therapy. Interestingly, our data do not support an effect of IVF therapies on serum LBP levels of participants, and by inference change in gut permeability yet up until 12 weeks. More research is needed to elucidate whether LBP can act as a surrogate marker of LPS and its related impact on inflammation and gut microflora. Additionally, dietary intake is one of the important modulators of intestinal microflora diversity and richness;[31] such information was not collected in the study. While it is nowadays common to link gut microflora impairment with the pathogenesis of certain metabolic disorders (such as obesity and T2D), which are also associated with low-grade inflammation, it is still however not yet confirmed whether inflammation is the cause or a consequence of the condition.[32]

Adiponectin is another useful marker of inflammation and has glucose-lowering properties.[17] Serum adiponectin inversely associates with BMI, fasting glucose and insulin, and TG levels, and positively associates with HDL-C levels.[17] These effects cannot be ascertained in our study, since interpretations are only based on unchanged levels of adiponectin within and between groups. In addition, the type of association between adiponectin and inflammation remains controversial in the literature. In contrast to the negative typical correlation between adiponectin, obesity and metabolic disease, a positive association was presented with inflammatory and immune-mediated diseases in one study.[33]

The interplay between inflammation and lipid metabolism is well documented, and the two play important role in the pathophysiology of metabolic conditions, such as insulin resistance.[34] Similarly to oral contraceptives, IVF hormones augmented lipid parameters, but values remained within normal range possibly due to the short duration of the treatment.[7,25] Metabolic and hormonal changes of pregnancy impose changes on lipid homeostasis, which play an important role in the provision of energy for the foetus. At 12 weeks of pregnancy, lipid profile complied with normal gestational changes with increased TG, T-Chol and HDL-C levels.[12] The attributable effect of IVF hormones on lipids in pregnancy cannot be however determined for the following reasons: (a) disparities in duration of IVF hormonal administration between the two groups (4 weeks for nonpregnant vs until 12 weeks of pregnancy); (b) a cumulative effect would have been identified if lipids parameters exceeded the reference ranges in pregnancy;[35] and (c) including a well-matched spontaneous-pregnant group would have allowed determination of the magnitude of change in lipid profile as an effect of pregnancy alone.

Baseline metabolic and endocrine parameters may differ between normal vs obese classified BMI, and this may have affected our data at 12 weeks. A limitation of this study hence relates to the large BMI range of participants. We did not perform subgrouping of BMI categories due to lack of power. Dietary intake, urine and stool sample collection of participants would have enabled more accurate assessment of early changes in gut microflora. Moreover, although the euglycaemic clamp protocol is considered as the 'gold standard' test for assessing changes in insulin sensitivity,[36] HOMA-IR is a surrogate measure, yet not as precise as the clamp but more practical and noninvasive for pregnant women. It is also worth mentioning that comparing the effect of IVF hormones to oral contraceptives is questionable, since the two therapies may differ in their type of bioactive oestrogen and progesterone and the duration of the treatment. In addition, although IVF therapies constitute much higher doses but they present a transient term of exposure to these hormones. In addition, pregnancy is a complex condition with significant inherent confounding effects on metabolic and inflammatory systems, which makes the direct comparison between pregnant and nonpregnant groups delicate. Finally, we acknowledge that other potential confounding factors may have influenced our data; such factors include psychological status and diet of participants. A well-matched spontaneous-pregnant group would have allowed the determination of magnitude of change in all these parameters as an effect of pregnancy alone. Future studies may want to compare more thoroughly pregnancies conceived spontaneously vs by IVF.