Disease condition and impact on patients Progesterone is a steroid hormone that is mainly formed in the cells of the corpus luteum of the ovary and during pregnancy in the placenta. Progesterone concentrations fluctuate during the menstrual cycle with levels barely detectable in the follicular phase and then rising with the increased synthesis that occurs after ovulation in the luteal phase. A surge in progesterone occurs approximately 1 day before ovulation. Progesterone increases the glandular component of the lining of the uterus (endometrium) in preparation to implant a fertilized ovum; during pregnancy, progesterone inhibits uterine contractions. Progesterone, in synergy with estrogen, also affects breast tissue by promoting the proliferation of the alveoli of the mammary gland. Infertility. Anovulation may be related to subfertility, PCOS, genetic disorders, obesity and other etiologies (1). PCOS is a complex disorder characterized by chronic anovulation and accounts for 70% of anovulatory infertility (2). Luteal phase deficiency (LPD). LPD corresponds to a duration of less than 11 days from ovulation until the onset of menses. Most experts believe it is a defect of corpus luteum progesterone output, both in amount and duration, that results in inadequate stimulation of the endometrium for implanting the fertilized ovum. It has also been implicated in early pregnancy loss. Does the test meet a medical need? Infertility. Although studies vary in support of serum progesterone levels to confirm ovulation, a single progesterone determination in the mid-luteal phase (approximately day 21 of the menstrual cycle or 7 days before anticipated menses) is frequently used (2). It has a reported sensitivity and specificity equal to or greater than 90% (3). A wide range of disorders can be considered once anovulation is confirmed, including hypogonadotropic and hypothalamic hypogonadism, premature or natural ovarian insufficiency, hyperprolactinaemia, thyroid dysfunction and late-onset congenital adrenal hyperplasia (4). All have a significant public health impact; but their diagnoses require more definitive testing than measuring progesterone levels. LPD. Because LPD as a cause of infertility has yet to be proven (5), interpreting results and applying them to patient management is challenging. But, given that progesterone is produced by the follicle, the follicle’s abnormal development could logically result in LPD. LPD may also impact fertility through an abnormal endometrial response to normal progesterone production. Despite the clear association of progesterone and the development of the corpus luteum, using progesterone to diagnose LPD is complicated by the pulsatile nature of progesterone secretion. Attempts to overcome this limitation by evaluating pooled luteal phase progesterone samples to improve sensitivity and specificity have been only somewhat successful (6). Nonetheless, clinicians have used empiric treatment with the understanding that evidence-based treatment regimens have not been developed. Further supporting the concept of LPD and its association with infertility, studies indicate that treatment with ovulation-inducing agents such as clomiphene or gonadotropins that optimize follicular development and number has been used with some success (7, 8) Progesterone levels have also been used to predict the outcome of early pregnancy when complications occur. Early diagnoses of a non-viable pregnancy allow consideration of management options that may help avoid unanticipated haemorrhage or pain (9, 10). Complications from spontaneous (and unsafe induced) abortions are recognized globally as a major public health concern. PCOS. Determining progesterone levels in women suspected of having PCOS but who are not amenorrhoeic can confirm anovulation and support the diagnosis. Appropriately diagnosing PCOS allows hormonal treatments that can regulate menstrual cycles, induce ovulation when indicated, decrease androgenic effects and also reduce the risks of metabolic abnormalities that often occur with PCOS (11). How the test is used Progesterone levels are primarily used in fertility investigations to confirm ovulation, evaluate PCOS and assess the proper functioning of the luteal phase. Apart from fertility evaluation, measuring progesterone may help establish lack of ovulation; but it cannot diagnose the wide range of conditions associated with anovulation, including hypothalamic and pituitary dysfunction as well as genetic disorders. Because the diagnosis of these disorders will depend on more definitive tests, this submission of progesterone as a diagnostic device focuses on disorders related to reproductive dysfunction. Infertility. Although there are many other tests that can signal ovulation (basal body temperature, cervical mucous, urinary LH testing), evaluating luteal phase progesterone level is often part of the basic workup of subfertile women to determine whether ovulation has occurred (1). LPD. LPD is diagnosed through a combination of assessments that may include determining progesterone levels, assessing basal body temperature, biopsy of the endometrium or sonographic imaging. Although progesterone measurement is clinically indicated in evaluating LPD, experts disagree on the cut-off point for abnormal levels, as well as the frequency and timing of testing. The accuracy of diagnosing LPD may be improved by evaluating mid-luteal progesterone levels on pooled samples. Progesterone can also be used to predict the outcome of early pregnancy in women experiencing complications such as pain or bleeding (12).

Prevalence and socioeconomic impact Infertility. Estimates suggest that up to 186 million women globally are infertile (13, 14). Although differing methods and definitions have been used to derive infertility burden, a recent review estimated global prevalence of infertility to be 9% (15). Data on worldwide prevalence of childlessness further estimate that as many as 70 million couples would benefit from medical intervention to achieve pregnancy (16). There are not enough data to assess global infertility prevalence trends over the past 20 years. But these trends can be impacted by a rise in STIs with subsequent impaired reproductive organ function, lifestyle changes and delayed childbearing (13–15). The economic and social impact of infertility is significant, particularly for women, who will often suffer from social isolation, discrimination, disinheritance, depression, abuse, divorce and possible abandonment in old age. Infertility as a common cause of childlessness can also have a broader negative economic impact on families, particularly in LMICs, where children contribute to family incomes and older parents depend on their children for support. PCOS is thought to be the most common endocrine disorder found in women of reproductive age and impacts all races and ethnicities. In unspecified populations PCOS has a reported incidence rate of 3–10% (17), although more precise incidence is unknown due to underdiagnosis. Common symptoms include irregular menstrual cycle, polycystic ovaries and hirsutism (18), but may also include infertility, insulin resistance, impaired glucose tolerance (type 2 diabetes) and dyslipidaemia (18). Multiple diagnostic criteria have been adopted for PCOS, but the common denominator appears to be oligoovulation and androgen excess. LPD. Estimates of LPD prevalence among infertile women vary considerably from around 4% to 30%, with variability attributed to lack of consensus on both its definition and diagnostic criteria. A shortened luteal phase is, however, reported to occur in 5% of ovulatory cycles and accounts for between 25% and 40% of recurrent pregnancy losses (19, 20).

There are no relevant guidelines for LPD. The 2006 guidelines by the Royal College of Obstetricians and Gynaecologists (12) considers whether serum progesterone assay has a role in predicting pregnancy outcome. They state that such an assay can be a useful adjunct when ultrasound suggests pregnancy of unknown location. In these cases, serum progesterone levels below 25 nmol/L are associated with pregnancies subsequently confirmed to be non-viable. But the guidelines warn against uterine evacuation based on a low initial progesterone, as viable pregnancies have been reported with initial levels less than 15.9 nmol/L. In the presence of pregnancy of unknown location, a serum progesterone less than 20 nmol/L predicts spontaneous pregnancy resolution with a sensitivity of 93% and specificity of 94%. One advantage is that the need for formal uterine evacuation can be reduced with a policy of expectant management. Levels above 25 nmol/L are “likely to indicate” and above 60 nmol/L are “strongly associated with” pregnancies subsequently shown to be normal. Overall, it is not possible to define a specific discriminatory value for a single serum progesterone result that will allow absolute clinical confirmation of viability or non-viability.

A 2012 review by Verhaegen et al. (21) involved 19 cohort studies, including more than 7000 women, and evaluated the diagnostic accuracy of a single serum progesterone measurement to predict pregnancy outcomes in women experiencing pain or bleeding with inconclusive ultrasound exam or with pain or bleeding alone. Sensitivity and specificity were highly dependent on the progesterone threshold selected. In women with pain or bleeding and an inconclusive ultrasound exam, a single progesterone measurement predicted a non-viable pregnancy with pooled sensitivity of 74.6% (95% CI: 50.6–89.4), specificity of 98.4% (95% CI: 90.9–99.7), positive likelihood ratio of 45 (7.1–289) and negative likelihood ratio of 0.26 (0.12–0.57). In women with pain or bleeding alone, progesterone predicted a non-viable pregnancy: • at a threshold of 10 ng/mL, with a pooled sensitivity of 66.5% (95% CI: 53.6–77.4), specificity of 96.3% (95% CI: 91.1–98.5), positive likelihood ratio of 18 (7.2–45), and negative likelihood ratio of 0.35 (0.24–0.50); • at a threshold of 15 ng/mL, with a pooled sensitivity of 83.3% (95% CI: 66.6–92.6), specificity of 87.5% (95% CI: 78.5–93.1), positive likelihood ratio of 6.7 (3.8–12) and negative likelihood ratio of 0.35 (0.09–0.5); and • at a threshold of 20 ng/mL, with pooled sensitivity of 85.7% (95% CI: 72.3–93.2), specificity of 66.6% (95% CI: 47–91.8), positive likelihood ratio of 2.6 (1.5–4.5) and negative likelihood ratio of 0.22 (0.1–0.47). These findings show that a single progesterone measurement for women in early pregnancy presenting with bleeding or pain and inconclusive ultrasound assessments can rule out a viable pregnancy. In 1996, McCord et al. (22) looked at the accuracy of screening serum progesterone to diagnose ectopic pregnancy (EP) and to identify a cut-off value that provides the best compromise between test sensitivity and specificity. The authors collected single progesterone measurements from 3674 pregnancies; they defined outcomes as EP, viable intrauterine pregnancy (IUP) and spontaneous abortion (SAB), and analysed the diagnostic accuracy of the test. They found that the diagnostic accuracy for: • EP vs IUP was 88.7%, standard error of the mean (SEM) = 0.1%; • SAB vs IUP was 93.8%, SEM = 0.4%; and • SAB plus EP vs IUP was 92.8%, SEM = 0.4%. They concluded that for progesterone > to 17.5 ng/mL, patients thought to be at risk for EP may be followed reasonably without ultrasound or further invasive diagnostic studies.

A 2011 review by Van der Linden et al. (23) describes various treatments during IVF for presumed LPD and concludes that progesterone supplementation seems to be an important aspect of any assisted reproductive treatment (ART). Vaisbuch et al. (24) surveyed clinicians providing ART in 2014 and reported that all 408 centres across 82 countries used some form of progesterone for luteal phase support. Since evidence-based data are lacking, a review by Mesen et al. in 2015 (25) discusses the controversy over whether LPD is a diagnosable disorder that is proven to cause infertility. The authors note that there are dual theories of LPD etiology (low levels of progesterone with a poorly functioning corpus luteum or a deficient endometrial response to normal hormonal levels), which may explain the lack of specific data to define LPD or to prove its association with infertility. The pulsatile nature of progesterone secretion also makes it hard to determine normal from subnormal levels. Although the review did not report a sensitivity or specificity for detecting LPD by progesterone level, a large study in normally menstruating women was cited where low progesterone levels were found to be significant for LPD by univariate analysis. The authors concluded that, despite the lack of evidence-based data, progesterone is critical for reproduction and LPD is a plausible cause of infertility and early pregnancy loss. A prospective study by Schliep et al. in 2014 (26) followed 259 women 18–44 years of age for up to two menstrual cycles and concluded that identifying ovulation in combination with a well-timed luteal progesterone measurement may serve as a cost-effective and specific tool for clinicians and researchers to assess LPD.

Progesterone measurement is an inexpensive test that can confirm ovulation as part of a basic infertility workup. Progesterone levels, despite the lack of evidence-based data, are used to support an LPD diagnosis, after which supplemental progesterone can be administered as a cost-effective way of facilitating conception. Supplementing ART cycles with progesterone can also decrease the cost of ART by increasing the success rate. Using a single progesterone measurement to predict a non-viable pregnancy in women with bleeding or pain may be particularly useful and cost-effective in LMICs where ultrasound equipment is not readily available. Timely diagnosis of a non-viable pregnancy allows planning expectant or medical management as needed and will decrease the number of complications and associated costs.

Progesterone levels are used to diagnose and manage infertility, which may not be a high priority in LMIC health care systems. In 2001, WHO recommended viewing infertility as a world health problem and encouraged the development of lower-cost ARTs. Progesterone measurement can be used towards this goal and can ultimately reduce inequalities in health care as it relates to infertility and childlessness. Using progesterone to identify non-viable pregnancies gives women the option of medical management even where ultrasound examination and other testing are not available.