The inadequate corpus luteum.

It is fascinating to look at follicular oestradiol production and luteal progesterone production on the same scale ( Fig. 1 ). Progesterone is the hormone that dominates the menstrual cycle. Progesterone concentrations above 25 nmol/L (7.7 ng/mL) are used in clinical practice to confirm ovulation. In a detailed monitored natural cycle, the average mid-luteal progesterone was 41.3 ± 3.2 nmol/L (13.0 ± 1.0 ng/mL) ( Groome et al. 1996 ) although normal pregnancies have been reported at less than half this level ( Yovich et al. 2015 ). In women undergoing gonadotrophin ovulation induction, the average timed mid-luteal progesterone concentrations in mono-ovulatory cycles was 49.6 nmol/L (15.6 ng/mL) with pregnancies reported at half this level ( Arce et al. 2011 ). Figure 1 Hormone profile of the menstrual cycle of women. Concentrations of oestradiol and progesterone plotted on the same scale highlighting the dominance of progesterone. Data adapted from Groome et al. (1996) , Duncan (2017) . Hormone profile of the menstrual cycle of women. Concentrations of oestradiol and progesterone plotted on the same scale highlighting the dominance of progesterone. Data adapted from Groome et al. (1996) , Duncan (2017) . In a study of endometrial secretory changes in the luteal phase, there was no evidence that infertile women had less decidualisation or more out of phase biopsies than fertile women. Indeed, there was significantly more abnormal luteal phase biopsies in the fertile population ( Filicori et al. 1984 ). In a study looking at artificial cycles, where intramuscular progesterone was used, low-dose progesterone that increased serum progesterone to 17.5 ± 3.5 nmol/L (5.5 ±1.1 mg/mL) was just as able to create normal secretory endometrial changes as high-dose progesterone (61.1 ± 21 nmol/L; 19.2 ± 6.6 ng/mL). The corpus luteum is a robust endocrine gland that produces more progesterone than is required for fertility. That makes sense as mutations or genetic variants that absolutely prevent conception are evolutionary dead ends and heavily selected against.

As luteal progesterone is absolutely required to support an early pregnancy until the luteoplacental shift, and removal of progesterone induces a miscarriage ( Csapo & Pulkkinen 1978 ), inadequate progesterone in early pregnancy would increase the risk of miscarriage and progesterone supplementation would mitigate this risk. There is, however, no evidence for low endogenous progesterone causing miscarriage. The carefully timed mid-luteal progesterone in 192 women undergoing ovulation induction with hCG trigger who went on to a full-term singleton delivery was 25.85 ± 10.08 ng/mL while those having a miscarriage was 28.64 ± 16.96 ng/mL ( Sallam et al. 1999 ). However, serum progesterone concentrations are lower in pregnancies that miscarry than in viable pregnancies ( Szekeres-Bartho & Balasch 2008 ). The hCG dynamics of a pregnancy that will miscarry are often abnormal with low progesterone concentrations a consequence of suboptimal hCG increases, in a failing pregnancy, rather than a direct cause of miscarriage. Although 15–20% of early pregnancies will miscarry in the first trimester the majority of these have chromosomal or morphological abnormalities and progesterone supplementation would have no benefit ( Chetty & Duncan 2018 ). Around 1% of couples suffer from recurrent pregnancy loss, here defined as three or more miscarriages. As this is higher than chance alone, there must be some underlying causes that predispose couples to miscarriage ( Chetty & Duncan 2018 ). While there may be genetic, structural or immunological causes most couples with recurrent miscarriage have no defined cause identified. Importantly, there is no reduction in luteal progesterone concentrations in women suffering from recurrent miscarriage ( Ogasawara et al. 1997 ). It may be that progesterone action is suboptimal in the presence of normal progesterone concentrations as it has been reported that decidualisation may be abnormal, with endometrial asynchrony, in women with recurrent miscarriage ( Szekeres-Bartho & Balasch 2008 ) suggesting an endometrial resistance to progesterone. However, similar findings have been reported in fertile women ( Filicori et al. 1984 , Coutifaris et al. 2004 ) and subfertile women with endometriosis ( Bulun et al. 2006 ) or implantation failure ( Timeva et al. 2014 ). Recurrent pregnancy loss is not usually associated with subfertility. It is difficult to argue that a measurable progesterone deficit directly causes miscarriage. Women with threatened miscarriage, and those with recurrent miscarriage, are often keen for treatment, and the use of progesterone supplementation to prevent miscarriage in early pregnancy is very common. The use of progesterone as a treatment in early pregnancy continues past the luteoplacental shift, when the luteal progesterone is minimal. The PROMISE study, a large placebo-controlled trial looking at progesterone supplementation in women with recurrent pregnancy loss, from the positive pregnancy test until 12 weeks of gestation, did not show evidence of efficacy for progesterone supplementation ( Coomarasamy et al. 2015 ). Reassuringly, however, women with six or more miscarriages had a more than 50% change of a normal pregnancy on the placebo arm of the study. Similarly, the PRISM study looking at progesterone supplementation in women with threatened miscarriage, until 16 weeks of gestation, also found no evidence of efficacy ( Coomarasamy et al. 2019 ). However a detailed analysis of the evidence suggests that if the woman has had a previous miscarriage or previous miscarriages that there was evidence of benefit from progesterone supplementation from onset of the bleeding ( Coomarasamy et al. 2020 ). What might be the effect of prolonged progesterone supplementation in women with threatened miscarriage where there is no evidence of suboptimal progesterone concentrations? It is likely that this is a pharmacological effect rather than a physiological replacement. Progesterone has anti-inflammatory effects, it is immune modulating and it causes quiescence of the myometrium ( Shah et al. 2019 ). It could be postulated that some women have an inflammatory or immune defect that predisposes to bleeding in early pregnancy and related increased uterine activity in response to the bleeding. Progesterone in supra-physiological concentrations may help alleviate this in some cases. This positive finding is likely to cause a major shift in how we manage women with threatened miscarriage and the expectation of women with early pregnancy bleeding. A common argument is that if it might benefit some women why not give it to all women with threatened or recurrent miscarriage. The safety data for progesterone in early pregnancy has been largely reassuring ( Piette 2020 ). There have been reports of an association of progesterone treatment in early pregnancy with developmental abnormalities such as hypospadias ( Carmichael et al. 2005 ). As this was not seen in the recent studies ( Coomarasamy et al. 2020 ) it might be an association with synthetic gestagen used in the past rather than natural progesterone used currently. There is also some evidence in humans for a link between progesterone exposure and the development of autism spectrum disorder (ASD) in later life ( Davidovitch et al. 2018 ). Using a national registry of male births Davidovitch showed that IVF treatment was not associated with ASD. However, progesterone treatment in early pregnancy was associated with ASD (RR 1.51: CI 1.22–1.86, P  < 0.001). Although correlation cannot lead to conclusions about causation, this is effect is biologically plausible as foetal steroids regulate epigenetic modulation of the brain, which has been postulated to be involved in the development of ASD ( Baron-Cohen et al. 2015 ). Studies on the sheep foetus, at the equivalent of 15 weeks of human gestation, after maternal administration of natural progesterone showed elevated concentrations in male but not female foetuses ( Siemienowicz et al. 2020 ). This was associated with functional changes in the pituitary gland and testes and an increase in circulating 11-dehydrocorticosterone, a steroid with mineralocorticoid effects ( Siemienowicz et al. 2020 ). Whether this translates into subtle changes in the male offspring phenotype is not known. However, it is well known that foetal exposure to an altered steroid environment has critical roles in sexual differentiation and the programming of health and disease in later life ( Ho et al. 2017 ). True luteal inadequacy does exist and it is iatrogenic in nature and there is a role for physiological progesterone supplementation in assisted conception. There also is a role for pharmacological progesterone supplementation in threatened miscarriage in women with previous miscarriages. However, we cannot say for certain that prolonged foetal exposure to increased progesterone beyond the luteoplacental shift has no effects on the adult offspring. This suggests that we should be cautious about the prolonged use of progesterone supplementation in early pregnancy outside the current guidelines or evidence base.

The author declares that there is no conflict of interest that could be perceived as prejudicing the impartiality of this commentary.