INTRODUCTION — Polycystic ovary syndrome (PCOS) accounts for the vast majority of anovulatory symptoms and hyperandrogenism in women [1]. The diagnosis of PCOS has lifelong implications, with increased risk for infertility, metabolic syndrome, type 2 diabetes mellitus, and, possibly, cardiovascular disease and endometrial carcinoma [2-5]. PCOS is diagnosed in adolescents with otherwise unexplained persistent hyperandrogenic anovulatory symptoms that are inappropriate for age and stage of adolescence. It should be considered in any adolescent girl with hirsutism, treatment-resistant acne, menstrual irregularity, or acanthosis nigricans, and evidence of these signs and symptoms should be especially sought in patients being evaluated for obesity. (See "Definition, clinical features, and differential diagnosis of polycystic ovary syndrome in adolescents", section on 'Clinical features'.)

This presentation of the etiology and pathophysiology of PCOS provides the rationale for the diagnosis and management of PCOS in adolescents, which are discussed separately:

●(See "Definition, clinical features, and differential diagnosis of polycystic ovary syndrome in adolescents".)

●(See "Diagnostic evaluation of polycystic ovary syndrome in adolescents".)

●(See "Treatment of polycystic ovary syndrome in adolescents".)

ETIOLOGY — PCOS is considered to be a complex trait arising from the interaction of genetic and environmental factors, usually first presenting when mature gonadotropin levels are achieved at puberty [1]. It occurs naturally in nonhuman primates as well as humans [6]. The pathogenesis of PCOS can be envisioned according to a "two-hit" hypothesis, whereby the disorder arises as a congenitally programmed predisposition ("first hit") that becomes manifest in the presence of a provocative factor ("second hit") (table 1) [1]. The congenital factors can be either hereditary (genetic) or acquired (eg, maternal androgens or nutritional disorders affecting the fetus). The postnatal provocative factor usually is insulin-resistant hyperinsulinism, which may have been congenitally programmed and/or acquired postnatally due to simple (exogenous) obesity.

These complex interactions generally mimic an autosomal dominant pattern of inheritance with variable penetrance. Heritability of PCOS has been estimated at over 70 percent, based on studies in identical twin sisters [7].

PCOS represents an evolutionary paradox: Although it is an infertility disorder, it is very common across populations [1]. One reasonable explanation for this paradox is that PCOS developed through natural selection because it was advantageous during times of nutritional deprivation, since the condition helps to preserve anabolism and reproductive capacity via increased androgen and insulin production and fat storage [8]. By contrast, in current times of plentiful nutrition and obesity, this phenotype is disadvantageous. One of the great mysteries about the etiology of PCOS is the nature of the common denominator that underlies ovarian hyperandrogenism, insulin resistance, and obesity. Considerable basic research will be necessary to discern this. Understanding of the pathophysiology is summarized in the following sections [1].

Heritable traits — Heritable traits that are PCOS risk factors include maternal PCOS, polycystic ovary morphology (PCOM), hyperandrogenemia, and metabolic syndrome.

Maternal polycystic ovary syndrome — Maternal PCOS is a risk factor for PCOS in daughters; approximately 25 percent of females with PCOS have a mother with PCOS, although estimates vary widely [9-11]. The largest long-term prospective study compared the daughters of 99 women with PCOS (defined by National Institutes of Health [NIH] criteria) with those of 88 control women [12,13]. Higher dehydroepiandrosterone sulfate (DHEAS) levels emerged peripubertally, and higher basal testosterone and 17-hydroxyprogesterone (17OHP) responses to gonadotropin-releasing hormone agonist (GnRHa) testing emerged during the late stages of puberty. One-half of the daughters of women with PCOS who were postmenarcheal had higher testosterone levels than any daughters in a control group. Twenty-one of these daughters were reported in 2019 to have reached 18 to 21 years of age: 11 met NIH criteria for PCOS (Rotterdam phenotypes A-B) and five had ovulatory PCOS (Rotterdam phenotype C) [14]. This represented a fivefold increased risk of developing PCOS compared with daughters of non-PCOS controls.

Polycystic ovary morphology — Polycystic ovary morphology (PCOM) refers to the ultrasonographic polycystic changes in the ovary that are variably associated with PCOS (see "Diagnostic evaluation of polycystic ovary syndrome in adolescents", section on 'Ultrasonography'). This characteristic seems to be inherited in an autosomal dominant fashion in PCOS [15,16]. In our study of 35 families with an adolescent PCOS proband (by NIH criteria), most adolescents had either a mother with PCOM (22 percent), usually without PCOS symptoms, or a parent with metabolic syndrome [9] (see 'Metabolic syndrome' below). In a study of the sisters of women with PCOM who met the broad Rotterdam criteria for PCOS (80 percent of whom were clinically hyperandrogenic), 70.5 percent of sisters had PCOM [17]. Among the sisters with PCOM, 25 percent had hyperandrogenic anovulation (ie, PCOS by the highly specific NIH criteria) (see "Definition, clinical features, and differential diagnosis of polycystic ovary syndrome in adolescents"), 42 percent had either hirsutism or oligomenorrhea (ie, meeting less specific criteria for PCOS), and 33 percent were asymptomatic. The group of sisters with PCOM resembled their proband sisters in both androgenic and glucose-metabolic traits, indicating significant heritability and cosegregation of these traits.

These and other genetic studies do not include data on the steroidogenic function of a polycystic ovary, which varies substantially among individuals with PCOM. Approximately one-half of apparently normal females with PCOM have subclinical biochemical evidence of ovarian androgenic dysfunction, which has been proposed to represent a carrier state for PCOS [1].

Hyperandrogenemia — Serum testosterone and sex hormone-binding globulin levels are heritable in women [18]. Approximately one-half of sisters of PCOS probands have an elevated serum testosterone level [19]. However, only one-half of these sisters with excess androgen have menstrual irregularity and the other one-half are asymptomatic. This is consistent with the concept that a modest degree of hyperandrogenemia is not sufficient to account for ovarian dysfunction in the absence of another predisposing or precipitating factor(s).

First-degree male relatives of women with PCOS have variably been reported to have subtle excesses of adrenal steroids and serum luteinizing hormone (LH) [20,21].

Metabolic syndrome — Metabolic syndrome (a cluster of central obesity, hyperglycemia, dyslipidemia, and hypertension) and its core pathogenetic constituents, insulin resistance and obesity, have heritable components [1] (see "Metabolic syndrome (insulin resistance syndrome or syndrome X)"). Defective insulin secretion is also highly heritable in PCOS and is closely associated with the most common subtype of PCOS [1,22]. (See 'Primary functional ovarian hyperandrogenism' below.)

There is a high prevalence of metabolic syndrome or its components (eg, insulin resistance and/or obesity) in first-degree relatives of PCOS patients, suggesting that these factors are strongly associated with the pathogenesis of PCOS [1,21]. This was illustrated in our series of 35 families in which an adolescent girl had PCOS (by NIH criteria): 70 percent of the probands had a parent with metabolic syndrome (53 to 79 percent of fathers and 34 to 37 percent of mothers were affected, depending on the criteria used) [9]. Obesity or overweight was found in 94 percent of fathers and 66 percent of mothers. Abnormal glucose tolerance was present in 84 percent of fathers and 49 percent of mothers; most of these were asymptomatic. In this series, there was complete concordance between paternal metabolic syndrome and PCOM in their proband daughter but no relationship to metabolic syndrome in these daughters; this suggests a fundamental relationship of metabolic syndrome to the PCOM of PCOS.

Genetic variation — Ongoing research aims to identify specific genes that underlie the intrinsic cause of PCOS. A wide variety of genetic variants with linkage to and/or association with PCOS have been identified by candidate gene and molecular genetic studies [1], and novel loci are reported regularly [23-27]. Many of these genes encode proteins with plausible roles in PCOS pathophysiology (eg, steroidogenic enzymes, sex hormone-binding globulin, androgen receptor, transcription factors, and gonadotropins and their receptors) or proteins related to insulin sensitivity and obesity.

The candidate gene approach led to the discovery of coding and regulatory variants in the AMH (anti-müllerian hormone) and AMHR (AMH receptor) genes in 6.7 percent of PCOS patients [28]. In silico and in vitro studies suggest that these variants cause decreased signaling and may have pathogenic significance. Linkages for polymorphisms in diverse signaling pathway genes have been robust in many populations [1]. Dysregulation of genes involved in cell growth or division is suggested by reports of upregulation of proto-oncogenic genes in endometrium and of telomere shortening in leukocytes of women with PCOS [1]. In most instances, it remains unclear how these genetic variants might interact to induce the core abnormalities of PCOS. The multiple pathways involved and lack of a common thread are, however, consistent with the observed heterogeneity of PCOS phenotypes and the concept that the basis of PCOS is multifactorial. (See "Epidemiology, phenotype, and genetics of the polycystic ovary syndrome in adults", section on 'Current perspective'.)

Several important investigations utilized genome-wide association studies (GWAS) to identify genes linked to PCOS, commencing with a series of studies conducted in large populations of ethnic Han Chinese in 2011 [1]. The major findings in this population have been generally replicated in most other populations studied. A meta-analysis of GWAS studies identified 13 loci common to all PCOS diagnostic criteria that were also associated with hyperandrogenism, gonadotropin levels, and testosterone levels, as well as the metabolic traits associated with PCOS [29]. Mendelian randomization analyses suggested that a shared genetic architecture contributes to PCOS, body mass index, fasting insulin, menopause timing, depression, and male-pattern balding. The molecular mechanisms by which most of the candidate genes plausibly contribute to PCOS are not clearly understood [30].

One of the most striking findings from GWAS studies in diverse PCOS cohorts has been the discovery of the DENND1A (differentially expressed in normal and neoplastic development isoform 1A, variant 2) locus [1,31-33]. The truncated splice variant of DENND1A, termed DENND1A.V2, encodes a previously unrecognized protein. DENND1A.V2 is normally found in theca and in zona reticularis cells and is overexpressed in PCOS theca cells, where it accounts for the PCOS biochemical phenotype in cultured theca cells [1,34]. In addition, in transgenic mice, overexpression of human DENNDIA.V2 drives both adrenal and ovarian steroidogenesis [35]. Several micro-RNAs have been incriminated in PCOS steroidogenesis and/or folliculogenesis [36-38]. These molecules inhibit gene expression via one or more post-transcriptional mechanisms. Theca cell profiling of micro-RNAs identified miR-130b-3p as one differentially expressed between women with PCOS and controls [30]. Expression of this micro-RNA was decreased in PCOS theca cells inversely to both DENND1A.V2 expression and androgen biosynthesis. Pathway and network analysis suggested that miR-130b-3p was at the hub of a network that involves DENND1A.V2 and two other GWAS candidate genes, RAB5B (Ras-related protein 5B, which encodes a GTPase involved in vesicular transport from the plasma membrane) and the gene encoding the LH receptor. miR-130b-3p also links this network to the insulin mitogenic signaling pathway. Thus, it was postulated that decreased miR-130b-3p may mediate the increased theca cell hyperandrogenism of PCOS by post-transcriptional upregulation of LH receptor signaling via this gene network [30]. DENND1A.V2 also may directly stimulate steroidogenic transcription [39].

A whole-genome sequencing study of 261 individuals from 62 PCOS families also implicated rare (minor allele frequency ≤2 percent) noncoding variants at the DENND1A locus in a number of PCOS-related reproductive and metabolic phenotypes [40]. One-half of these PCOS families had one or more of 30 rare DENND1A variants. The vast majority of these were noncoding variants, most of which were predicted to alter V2 expression via altered binding of transcription factors or RNA-binding proteins; it was proposed that these variants regulate key biologic pathways of DENND1A action. (See "Epidemiology, phenotype, and genetics of the polycystic ovary syndrome in adults", section on 'Genetics'.)

Epigenetic changes — Epigenetic changes are increasingly suspected to be important to the pathogenesis of PCOS. Abnormal epigenetic methylation of cytosine-guanine (CpG) islands have been identified in tissues from adult women with PCOS. These alter gene expression and may reflect environmental exposures [41]. Methylation differences at the androgen receptor (AR) intron-exon boundaries were observed that were postulated to promote AR alternative splicing in PCOS granulosa cells [42]. Two novel splice variants of the AR were discovered in these cells that altered both AR recruitment to DNA and androgen-mediated changes in the expression of genes affecting steroidogenesis and folliculogenesis [42]. A comprehensive epigenomic study of granulosa lutein cells from women with PCOS compared with controls used an 850k methylation array and identified >100 differentially methylated CpG sites associated with 88 genes; pathway analyses suggested a disruption of pathways related to cancer, cardiogenesis, Hedgehog signaling, and the immune response [43]. Abnormal DNA methylation has been observed in multiple PCOS tissues [44].

Intrauterine environment — Insults to the intrauterine environment induce persistent changes in the developing embryonic and fetal epigenome that lead to altered gene expression and disease in later life [1]. Excess prenatal androgen exposure and intrauterine over- and undernutrition have been incriminated as risk factors for PCOS.

Congenital virilization — PCOS is often observed in adolescents with congenital virilizing disorders, including congenital adrenal hyperplasia (CAH) [1]. In untreated nonclassic CAH patients, hyperandrogenism is accompanied by significant insulin resistance [45,46].

Postnatal clinical markers of in utero androgen excess, such as anogenital distance and the ratio of the second-to-fourth digits of the hand (2D:4D), have been explored in offspring of women with PCOS, with mixed results [1,47,48].

Prenatally androgenized animal models of PCOS, on the other hand, have shown androgen to be a potent epigenetic programmer during embryonic/fetal development in many tissues and in numerous species; these effects extend well beyond the established role of androgen in masculinizing the genitals [49]. Studies in nonhuman primates exposed to androgen excess early in gestation have yielded a model closely resembling PCOS. The critical window for androgen programming in rhesus monkeys is during early to mid-gestation [1,49]. Prenatally androgenized female rhesus monkeys are born small and fat after experiencing compromised placental function and exhibit exaggerated weight gain during infancy [50]. They then develop the characteristic reproductive features of PCOS, including ovarian and adrenal hyperandrogenism, oligomenorrhea, polyfollicular ovaries, elevated LH levels, and resistance to progesterone negative feedback on GnRH/LH secretion. They also have PCOS-like metabolic abnormalities, including truncal obesity, insulin resistance, impaired glucose tolerance, and dyslipidemia.

Studies in prenatally androgenized rodent models of PCOS indicate that prenatal androgens suppress the induction of hypothalamic progesterone receptors (PR) by estrogen; hypothalamic PR suppression persists postnatally and is associated with increased LH levels and pulsatility, absent LH surges, disrupted estrous cycles, and hyperandrogenemia in adulthood [51-53]. These mice also developed a specific increase of GABAergic synapsing on GnRH neurons before puberty, and antiandrogen administration in adult mice restored both normal central GABAergic wiring and reproductive function [53]. While the mechanism remains to be determined, this is a remarkable indication of central nervous system plasticity and reversibility of a PCOS-like state.

The proposed role of epigenetics in PCOS has been bolstered by the observation that the prenatal effect of virilization is transgenerational in mice [14]. Female mice with PCOS-like traits (with and without obesity) induced by prenatal androgenization were found to produce female offspring (the F₃ generation) with PCOS-like reproductive and metabolic traits. Common gene expression signatures were found in oocytes of F₃ mice and in blood cells of PCOS daughters and unrelated women with PCOS, including upregulation of the gene encoding TIAL1, an RNA-binding protein.

These observations demonstrate the potential for in utero exposures to permanently alter postnatal function via epigenetic reprogramming, while giving an early indication of plasticity of the peripubertal reproductive axis. While prenatal testosterone administration has proven to be a useful experimental tool for inducing a PCOS-phenotype in animal models, it is highly unlikely that intrauterine androgen excess either from the mother or from exogenous exposures underlies PCOS, due to the high expression of aromatase by the human placenta. The fetal ovary is also unlikely to be a source of excess androgen since it develops in the second trimester and is functionally quiescent [1]. Congenital virilization in other model systems suggest that it is not increased androgens or AR activity per se, but rather a change in the level of one or more intermediates in the post-receptor androgen signaling pathway (eg, prostaglandins), which intersect with other effector pathways (eg, glutamate receptor signaling) to influence masculinization [54,55]; in a neonatal brain virilization model, this has involved epigenetic alteration [56]. Something similar may underlie the pathogenesis of PCOS.

Disturbed fetal nutrition — There is good evidence that fetal undernutrition predisposes to metabolic syndrome and related cardiovascular disease in adulthood; this association is sometimes termed the "Barker hypothesis," "fetal origins of adult disease," or "metabolic programming." It has been proposed that low birth weight is likewise a risk factor for PCOS [57]. This proposal has been supported by some but not all studies [1]. In some studies, high birth weight has also been associated with PCOS. However, most cases of PCOS in most populations occur in individuals with normal birth weight [58,59]. (See "Definition, epidemiology, and etiology of obesity in children and adolescents", section on 'Metabolic programming'.)

Postnatal environment — Postnatal environmental risk factors can be viewed as a second "hit," which cause a latent heritable or congenitally programmed susceptibility trait to manifest as PCOS (table 1).

Insulin-resistant hyperinsulinism — All states of extreme insulin resistance (eg, insulin receptor loss-of-function mutations and generalized lipodystrophy) are associated with PCOS [1]. Pseudo-Cushing's syndrome and pseudoacromegaly are two rare syndromes of intractable pediatric obesity that herald PCOS in adolescence. They are both characterized by moderately severe insulin resistance and compensatory hyperinsulinemia that cause obesity and overgrowth, respectively. (See 'Insulin-resistant hyperinsulinism' below and "Causes and pathophysiology of Cushing's syndrome", section on 'Pseudo-Cushing's syndrome' and "Insulin resistance: Definition and clinical spectrum", section on 'Clinical features'.)

Simple (exogenous) obesity is the most common cause of insulin resistance. Insulin resistance causes compensatory hyperinsulinemia, which promotes anovulation in women with PCOS (see 'A unified concept of PCOS pathophysiology' below). Accordingly, ovulatory patients with PCOS are more insulin-sensitive than are anovulatory patients with PCOS [1,60]. Any medical treatment or lifestyle modification that improves indices of insulin sensitivity in patients with PCOS also improves menstrual cyclicity and ovulatory function [1]. (See "Treatment of polycystic ovary syndrome in adolescents".)

Hyperandrogenism — In girls with poorly controlled congenital adrenal virilizing disorders such as CAH, the androgen excess causes ovarian hyperandrogenism and PCOM that is generally reversible when androgen levels are controlled with glucocorticoids [61]; this is consistent with the PCOS-like state induced by peripubertal virilization of animal models that is reversible upon withdrawal of androgen, as discussed next [49]. In some girls, however, ovarian hyperandrogenism persists, which seems to be due to the epigenetic reprogramming by congenital virilization discussed above.

Peripubertal virilization of nonhuman primates and rodents causes a PCOS-like state that is reversible upon withdrawal of androgen [49]. Studies in transgenic mice suggest that the brain is the key target: In peripubertally virilized mice, selective ablation of the AR in the brain completely or partially rescued the PCOS-like phenotype, whereas silencing of the AR in granulosa cells was not protective [62].

Premature adrenarche seems to pose approximately a twofold increased risk for developing PCOS [1,57]. The nature of this association is unclear. We favor the concept that premature adrenarche is sometimes an early manifestation of the steroidogenic dysregulation that underlies PCOS, though in some populations, low birth weight has been postulated to be the primary risk factor underlying both premature adrenarche and PCOS. (See "Premature adrenarche".)

Other precipitants and risk factors — The best support for a role of LH excess is in the setting of congenital virilization, in which there is prenatal reprogramming that may underpin LH excess at puberty. (See 'Congenital virilization' above.)

Endocrine disruptors such as bisphenol A have been suspected of aggravating PCOS. While bisphenol A levels are slightly higher in individuals with PCOS compared with controls, there is no convincing human evidence that these minor differences are causally related to the development of PCOS [63,64]. (See "Occupational and environmental risks to reproduction in females: Specific exposures and impact", section on 'Bisphenol A and other phenols'.)

A UNIFIED CONCEPT OF PCOS PATHOPHYSIOLOGY — Approximately 90 percent of individuals with PCOS have abnormal ovarian androgenic function [1,22,65]. An abnormal degree of insulin resistance, obesity, or luteinizing hormone (LH) excess are each found in approximately one-half of cases. Thus, the common denominator in PCOS appears to be ovarian hyperandrogenism, with insulin-resistant hyperinsulinism being a nonessential but common aggravating factor in the pathophysiology. LH excess and the propensity to obesity seem to be secondary to the underlying ovarian hyperandrogenism and hyperinsulinism.

We favor a unified minimal model of PCOS pathophysiology, which incorporates the major features of the syndrome into a parsimonious scheme [1] and is shown in the figure (figure 1). A brief overview:

●Functional ovarian hyperandrogenism (FOH) – The model posits that FOH, which is found in the vast majority of cases, is the essence of PCOS. FOH can account for the essential clinical features of PCOS: hirsutism, oligo-anovulation, and polycystic ovaries.

In most individuals with PCOS, the cause of ovarian hyperandrogenism seems to be intrinsic to the ovary. In a small minority of individuals, modest hyperandrogenemia of extra-ovarian origin (adrenal or peripheral sources) causes anovulation and polycystic ovaries. (See 'Secondary functional ovarian hyperandrogenism' below.)

●Hyperinsulinism and obesity – Approximately one-half of cases of PCOS have insulin resistance with compensatory hyperinsulinemia. This hyperinsulinemia exerts paradoxical effects because insulin activity is tissue-specific: Hyperinsulinemia affects the ovary because it remains insulin-sensitive, whereas skeletal muscle and the liver demonstrate insulin resistance (see 'Insulin-resistant hyperinsulinism' below). In the ovary, hyperinsulinism synergizes with LH to upregulate androgen production and also causes premature luteinization of granulosa cells. These actions of insulin aggravate hyperandrogenism, development of polycystic ovary morphology (PCOM), and anovulation (see 'Insulin-resistant hyperinsulinism' below). Insulin-resistant hyperinsulinemia also promotes adiposity, which exacerbates the insulin-resistant state. (See 'Associated pathophysiologic disturbances' below.)

●LH excess – The moderate hyperandrogenemia causes secondary LH elevation by interfering with progesterone negative feedback at the hypothalamic level. LH excess on its own is not likely to cause ovarian hyperandrogenism, because the ovary normally demonstrates homologous desensitization when exposed to high levels of LH. However, this protective mechanism is overridden in an environment of insulin excess (or increased DENNDA1.V2 expression), wherein the ovary becomes hyperresponsive to LH and produces more androgens.

These major features are discussed individually in the following sections.

Primary functional ovarian hyperandrogenism — Approximately 90 percent of PCOS patients have FOH, based on results of ovarian androgenic dynamic testing [1,22,65]. FOH can account for the essential features of PCOS ((figure 1), step 1). The excessive circulating androgens act on pilosebaceous units of the skin to cause the cutaneous manifestations of the syndrome, eg, hirsutism and acne. The excessive intraovarian androgenic milieu causes granulosa cell dysfunction, which manifests as oligo-anovulation and often PCOM. (See "Definition, clinical features, and differential diagnosis of polycystic ovary syndrome in adolescents", section on 'Clinical features'.)

PCOS can be divided into several functional categories, representing different sources of androgen excess (table 2 and figure 2):

●Functionally typical PCOS/FOH – Two-thirds of cases are considered "functionally typical" because they demonstrate an exaggerated increase in 17-hydroxyprogesterone (17OHP) in response to LH receptor stimulation achieved clinically with either a gonadotropin-releasing hormone agonist (GnRHa) or with human chorionic gonadotropin. The increased 17OHP response appears to be an indicator of dysregulated cytochrome P450c17 function, as discussed next. In addition, these patients have an abnormal response to the dexamethasone androgen-suppression test (DAST): Dexamethasone suppresses adrenocorticotropic hormone (ACTH)-dependent adrenocortical steroidogenesis but fails to suppress serum testosterone, consistent with an ovarian source. (See "Diagnostic evaluation of polycystic ovary syndrome in adolescents".)

●Functionally atypical PCOS – The remaining one-third of cases are considered "functionally atypical" because they have a normal 17OHP response to GnRHa or human chorionic gonadotropin but an abnormal response to DAST.

Some patients with functionally atypical PCOS have isolated primary functional adrenal hyperandrogenism (FAH), demonstrated by ACTH stimulation testing, or PCOS without either FOH or FAH. Most of the latter group have the atypical PCOS of obesity, while a minority have idiopathic atypical PCOS (table 2 and figure 2). Both groups are described below. (See 'Functional adrenal hyperandrogenism' below and 'Obesity' below.)

Mechanisms for FOH include:

●Dysregulation of theca cell steroidogenesis — The pattern of steroid secretion in PCOS typically indicates generalized overactivity of the entire steroidogenic pathway from cholesterol to androgens and estrogens, with the prominent elevation in 17OHP indicating dysregulation of cytochrome P450c17 activities, which are encoded by the CYP17A1 gene [1,66]. Cytochrome P450c17 has both 17-hydroxylase and 17,20 lyase activity, and the latter is the rate-limiting step in androgen formation.

Androgen production within the ovary is normally tightly regulated because androgen is a necessary evil in the ovary: Androgen is required as the substrate for estrogen formation, and androgen itself is required for optimal fertility [1,67]. However, excessive androgen disrupts ovarian follicular development. At the small antral follicle stage of development, LH stimulates ovarian theca cells to form androgen and follicle-stimulating hormone (FSH) stimulates granulosa cells to form estrogen (figure 3).

Normally, theca cell androgen production is coordinated with granulosa cell estrogen production by intraovarian (paracrine/autocrine) mechanisms rather than by endocrine negative feedback loops (figure 3). Within the ovary, the response to LH is modulated by the balance between downregulation and upregulation processes. Downregulation results from desensitization of LH receptor-binding sites that occurs in the presence of elevated LH levels ("homologous desensitization") and by inhibition of P450c17 activities by local androgens and estrogens. Downregulation is counterbalanced by granulosa cell factors that upregulate P450c17 activities, particularly insulin-like growth factors (IGFs) and inhibin. Extraovarian factors such as insulin and proinflammatory cytokines (eg, tumor necrosis factor [TNF]-alpha, transforming growth factor-beta, and interleukin-6) produced by specialized monocytic cells of adipose tissue and the gut microbiome also upregulate P450c17 activities, while other cytokines (eg, interleukin-22) downregulate androgenic enzymes [68,69].

In PCOS, downregulation of thecal androgen production is fundamentally flawed: PCOS ovaries are hyperresponsive to LH stimulation (figure 1) [1,66]. Three lines of evidence suggest that the dysregulation of steroidogenesis in PCOS is fundamentally of intrinsic ovarian origin [1]. First, clinical studies in women with PCOS have demonstrated that FOH steroidogenic hypersensitivity to gonadotropin stimulation persists after prolonged gonadotropin suppression with GnRHa; there is also a parallel primary adrenal steroidogenic defect in some women, termed primary FAH, as discussed in the following sections. Second, in vitro studies have shown that the abnormal steroidogenic phenotype that is typical of PCOS theca cells persists in long-term culture, consistent with an intrinsic theca cell genetic defect. Third, genome-wide association studies (GWAS) have identified genetic linkage of PCOS to the DENND1A.V2 protein that reproduces the PCOS theca cell phenotype in vitro, as discussed above. (See 'Genetic variation' above.)

●Granulosa cell dysfunction – Granulosa cells line the antrum, or central fluid-filled cavity, of the growing ovarian follicle. They form estrogens from androgens (via expression of aromatase) and secrete peptides and growth factors that regulate follicle and oocyte development (figure 3) [70]. Androgen actions on granulosa cells are required for optimal follicle growth in response to gonadotropins [1,67]. Androgens normally stimulate the transformation of primordial follicles from dormancy into primary follicles ("recruitment"), which is the first step in their growth and development into small antral follicles [71]. Androgens then normally synergize with FSH to induce LH receptor expression on granulosa cells; this initiates "luteinization," which permits granulosa cells to form progesterone in response to LH and begins the formation of the preovulatory follicle [1].

In PCOS, excess intraovarian androgen causes the granulosa cell dysfunction that underlies anovulation and PCOM (figure 1) [1]. High androgen levels stimulate an excessive number of small follicles to grow [1]. This stimulation of early follicle growth, primarily resulting from increased granulosa cell proliferation, is most likely responsible for the increased anti-müllerian hormone (AMH) levels in women with PCOS; AMH is a product of the granulosa cells of small growing follicles that normally restrains primordial follicle growth ("recruitment") (see "Diagnostic evaluation of polycystic ovary syndrome in adolescents"). Androgens also contribute to the premature luteinization of the follicle, ie, granulosa cell luteinization occurs in the mid- rather than the late follicular phase, which seems to be the basis for the follicular maturation arrest that hinders the emergence of a dominant follicle and precludes ovulation [72]. The increased proliferation of luteinized small follicles coupled with maturation arrest before the large antral stage of follicle development manifests as PCOM on ultrasound (figure 1).

Alternatively, it is possible that granulosa cell dysfunction in PCOS is a manifestation of an intrinsic defect in the intraovarian (paracrine) regulation of follicle dynamics. For example, excessive granulosa cell secretion of factors such as inhibin B that stimulate androgens may reflect intrinsic granulosa cell dysfunction. Indeed, a subset of women with PCOS have deleterious variants in the genes encoding AMH or its receptor that impair restraint of follicular growth and granulosa cell aromatase activity; impaired AMH signaling has been postulated to underlie PCOS in these patients [28], but it remains to be determined whether these patients are also atypical in having evidence of the premature ovarian failure characteristic of Amh null mouse models [73].

Insulin-resistant hyperinsulinism — Hyperinsulinism is an aggravating factor in approximately one-half of PCOS cases; these women manifest excessive insulin resistance relative to the severity of their obesity (body mass index) [1,74,75]. Thus, in these women, insulin resistance is thought to be intrinsic. The other one-half of PCOS patients are no more insulin-resistant than otherwise healthy individuals with comparable body mass indexes.

Hyperinsulinism does not account for the intrinsic theca cell dysfunction characteristic of ordinary PCOS, because it is inconsistently present. Furthermore, the insulin resistance is usually mild, particularly in the developmental (adolescent) phase of the syndrome. Hyperinsulinism only plays a primary pathogenic role in the development of PCOS in patients with severe or extreme insulin-resistance syndromes.

The insulin resistance of PCOS is remarkable because resistance to the metabolic effects of insulin is tissue-specific [1]. While tissues such as skeletal muscle are resistant to insulin action as it relates to glucose metabolism, the compensatory hyperinsulinemia elicits excess insulin action in the ovary, liver, and adipocytes [76]. Signaling pathways remain sensitive to many growth factor-like, protein-anabolic, and lipogenic actions of insulin. These actions of insulin aggravate the steroidogenic dysregulation and contribute to the obesity and acanthosis nigricans that are common comorbidities of PCOS. Thus, PCOS presents the paradox of "insulin-resistant hyperinsulinism," ie, a state of excessive insulin signaling in the presence of resistance to the main glucoregulatory effects of insulin.

Hyperinsulinism, when present, is an important aggravating factor in PCOS pathogenesis because it promotes FOH: The hyperinsulinemia sensitizes the intrinsically dysregulated ovarian theca cells to secrete excess androgen in response to LH [1] ((figure 1), step 2). It does this by reversing the LH-induced homologous desensitization of LH-binding sites, which then upregulates cytochrome P450c17 activities (figure 3). Insulin also upregulates testosterone formation by type 5 17-beta-hydroxysteroid dehydrogenase via a transcription factor (KLF15) that stimulates adipogenesis [77]. Insulin excess also synergizes with androgen excess and FSH to prematurely induce LH receptors on granulosa cells, leading to their premature luteinization and follicle maturation arrest [1,72]. The severity of insulin resistance correlates with the severity of hyperandrogenism, severity of anovulatory symptomatology, and prevalence of PCOM. Any treatment that lowers insulin levels therefore improves hyperandrogenism (including in experimental models of PCOS [78]), though the hyperandrogenism of typical PCOS is seldom fully corrected by such treatments alone in most populations.

One of the great mysteries about PCOS is whether the hyperandrogenism might cause insulin resistance. Androgens in vitro exert some anti-insulin effects [79]. A variety of studies indicate that administration of oral androgens or virilizing doses of androgen to humans causes insulin resistance [79,80]. The induction of modest hyperandrogenism does not cause insulin resistance in women [81], but in adult nonhuman primates, it will induce insulin resistance when combined with an obesogenic diet [82,83]. However, reversal of PCOS androgen excess has not ameliorated insulin resistance in the majority of studies [84]. In animal models, androgens permanently program insulin resistance when given transiently prenatally. (See 'Congenital virilization' above.)

Associated pathophysiologic disturbances — Excess gonadotropin secretion, adiposity, and adrenal androgenic dysfunction are inconsistent features of PCOS that seem related to dysregulation of steroidogenesis and/or insulin-resistant hyperinsulinism.

Gonadotropin abnormalities — LH is necessary for the expression of steroidogenic enzymes. Thus, PCOS is a gonadotropin-dependent, ie, functional, form of ovarian hyperandrogenism. Any treatment that suppresses LH levels (eg, estrogen-progestin oral contraceptives) suppresses ovarian hyperandrogenism [1].

Increased LH, often accompanied by decreased FSH, was the first laboratory abnormality identified in classic PCOS and was historically thought to play a role in its pathogenesis by increasing ovarian androgen production. Subsequent research showed that serum LH is elevated in only approximately one-half of PCOS patients with documented FOH [1]. Furthermore, the homologous desensitization phenomenon limits ovarian androgen excess in response to excess LH, except in the presence of insulin-resistant hyperinsulinism.

Evidence suggests that the increased LH observed in patients with PCOS is primarily caused by androgen interference with progesterone negative feedback [1]. This resistance is overcome by antiandrogen treatment [85]. Resistance to estrogen-progesterone negative feedback, while significant, occurs in less than one-half of adolescents with PCOS, in contrast with almost all adults [86]. This discrepancy between adolescents and adults suggests that this resistance only becomes apparent when the high sensitivity to sex-steroid negative feedback of childhood fully wanes with maturity.

Such data suggest that LH excess in PCOS is the result, rather than the cause, of androgen excess.

In more severe cases of PCOS, the moderate testosterone excess stimulates LH secretion ((figure 1), step 3). Additionally, granulosa cells that have been prematurely luteinized, under the influence of both increased testosterone and insulin levels (see 'Insulin-resistant hyperinsulinism' above), begin to secrete estrogen in response to LH as well as to FSH. The negative feedback effect of the resultant estradiol excess can account for the significantly low FSH levels of PCOS ((figure 1), step 4).

Obesity — Approximately one-half of PCOS patients are obese, and at least one-third of nonobese PCOS patients have increased intraabdominal fat [1,87].

The reason for the high prevalence of obesity in PCOS is not entirely clear. The hyperinsulinism of insulin resistance seems to be an important factor since in vitro studies indicate that insulin signaling in human subcutaneous adipose tissue is intact in PCOS [1,88]. Insulin signaling is of major importance to the size and function of the white adipose tissue depot: It stimulates adipogenesis (differentiation of preadipocytes into adipocytes) and lipogenesis, while inhibiting lipolysis [88-90]. Adolescence is a critical period for establishment of the adipocyte population. Obesity may also arise to some extent from an intrinsic developmental defect in the adipocyte lineage [91]. Androgens oppose insulin effects on subcutaneous fat stores [8,46], while mild hyperandrogenemia promotes visceral fat accumulation in young female primates fed a high-fat diet [41,92]. There is also evidence that PCOS abdominal adipose stem cells exhibit a precocious commitment to a preadipocyte lineage, which ultimately leads to an increased number of adipocytes and lipid storage [93].

Obesity aggravates the clinical severity of FOH by increasing insulin resistance ((figure 1), panel B). The mechanisms by which obesity causes insulin resistance are unclear [94] but may include a deficiency of insulin-sensitizing adipokines such as adiponectin in favor of proinflammatory cytokines such as TNF-alpha that are secreted by a unique population of adipose tissue-derived regulatory T cells [95,96] and by an altered intestinal microbiome [68,69]. In PCOS, visceral fat contributes more to insulin resistance than does subcutaneous abdominal fat because of its enhanced lipolytic response to catecholamines (the major lipolytic stimulus in humans) [1]. Enhanced visceral fat lipolysis is unique to PCOS and not attributable to androgen excess [97,98]; the free fatty acids and diacylglycerol released from visceral fat appear to promote insulin resistance in the liver and muscle, respectively. (See "Pathogenesis of type 2 diabetes mellitus", section on 'Role of diet, obesity, and inflammation'.)

Inflammatory cytokines such as TNF-alpha that are produced by the mononuclear cells of excess white adipose tissue also aggravate ovarian hyperandrogenism. Hyperandrogenism sensitizes circulating mononuclear cells to secrete inflammatory cytokines in response to glucose and saturated fat ingestion [1,99]. In turn, androgen levels and insulin resistance increase in response to glucose or saturated fat loading [100,101].

In contrast with energy-storing white adipose tissue, women with PCOS have been shown to have less energy-consuming brown adipose tissue [102,103]. Study of the regulation of brown fat development is in its infancy [104].

In addition to aggravating PCOS via insulin resistance, obesity seems to be the sole cause of most functionally atypical PCOS (sometimes termed "atypical PCOS of obesity") (table 2 and figure 2) [1]. Both androgen and estrogen are formed in excess by the adipose tissue of obese women [77,105-107]. Obesity also suppresses gonadotropin levels by accelerating their metabolism [108,109]. Thus, obesity should be suspected as the cause of menstrual irregularity when dehydroepiandrosterone sulfate (DHEAS) and LH are normal in obese PCOS women with marginally elevated testosterone levels.

Functional adrenal hyperandrogenism — In 25 to 50 percent of cases, primary FOH is accompanied by a seemingly related primary FAH. In another 5 percent of cases of PCOS, FAH is the only detectable source of androgen excess (table 2 and figure 2 and table 3) [1]. Isolated primary FAH usually occurs in hirsute women without menstrual abnormalities, ie, without PCOS.

Primary FAH is defined as 17-ketosteroid hyperresponsiveness to ACTH that is otherwise unexplained. The steroidogenic pattern of response to ACTH resembles an exaggerated adrenarche and, in the past, was confused with nonclassic 3-beta-hydroxysteroid dehydrogenase deficiency (see "Definition, clinical features, and differential diagnosis of polycystic ovary syndrome in adolescents"). DHEA is the major hyperresponsive 17-ketosteroid and is correlated with 17-hydroxypregnenolone responses. This FAH has been postulated to result from a dysregulation of adrenal zona reticularis steroidogenesis that parallels the dysregulation of ovarian theca cell steroidogenesis [1,110]. This FAH has been associated with mild adrenal enlargement and a small degree of autonomous adrenocortical function in some cases [111]. Hyperinsulinism seems to aggravate FAH as it does FOH.

Secondary functional ovarian hyperandrogenism — Hyperandrogenic anovulation may be explained by a variety of relatively uncommon disorders (table 3) (see "Definition, clinical features, and differential diagnosis of polycystic ovary syndrome in adolescents"). Several such disorders are capable of causing secondary FOH.

Virilizing disorders, the most common of which is virilizing congenital adrenal hyperplasia (CAH), can cause hyperandrogenic anovulation with PCOM and FOH in a combination of ways [1,61]. The mechanisms include both congenital virilization, which programs for PCOS at puberty regardless of how well postnatal virilization is controlled (see 'Etiology' above), and uncontrolled hyperandrogenism, which causes PCOM and follicle maturation arrest. Rarely, patients with CAH have ectopic adrenal rest tissue, which is sensitive to LH and causes hyperandrogenism that cannot be controlled by adrenal suppression [61].

PCOS due to secondary FOH is also seen in any disorder that causes extreme insulin resistance and hyperinsulinemia, for example, insulin-receptor mutations (where extreme hyperinsulinemia activates the IGF-1 receptor) and generalized lipodystrophy [1]. In these cases, PCOS features may appear as early as the neonatal period or early childhood. A significant elevation of serum IGF-1 accounts for the secondary FOH of acromegaly [112].

Valproic acid, an antiseizure medication, directly augments the transcription of the steroidogenic gene CYP17 that encodes cytochrome P450c17 and can cause hyperandrogenic anovulation and PCOM. (See "Definition, clinical features, and differential diagnosis of polycystic ovary syndrome in adolescents", section on 'Differential diagnosis'.)

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Polycystic ovary syndrome" and "Society guideline links: Hirsutism".)

SUMMARY — Polycystic ovary syndrome (PCOS) is the most common cause of anovulatory symptoms and hyperandrogenism in women, and symptoms typically first manifest during adolescence.

●PCOS seems to arise as a complex trait that results from the interaction of diverse genetic and environmental factors that usually first manifest when mature gonadotropin levels are achieved at puberty. Obesity is the most common postnatal environmental contributor. (See 'Etiology' above.)

●The pathophysiology of PCOS usually includes excess ovarian androgen production (functional ovarian hyperandrogenism [FOH]), with or without insulin-resistant hyperinsulinism or functional adrenal hyperandrogenism (FAH) (figure 1 and figure 2). (See 'A unified concept of PCOS pathophysiology' above.)

•FOH is present in most cases and typically arises from dysregulation of steroidogenesis that includes an increased sensitivity to luteinizing hormone (LH) ((figure 1), step 1). (See 'Primary functional ovarian hyperandrogenism' above.)

•In approximately one-half of cases, PCOS is associated with a selective form of insulin resistance, in which muscle is resistant to the metabolic effects of insulin, while the ovaries, adrenals, and adipose tissue remain relatively sensitive to insulin, thereby promoting androgen production and obesity in response to compensatory hyperinsulinemia ((figure 1), step 2). (See 'Insulin-resistant hyperinsulinism' above.)