The Diagnosis of Polycystic Ovary Syndrome in Adolescents

Abstract

Consensus has recently been reached by international pediatric subspecialty societies that otherwise unexplained persistent hyperandrogenic anovulation using age- and stage-appropriate standards are appropriate diagnostic criteria for polycystic ovary syndrome (PCOS) in adolescents. The purpose of this review is to summarize these recommendations and discuss their basis and implications. Anovulation is indicated by abnormal uterine bleeding, which exists when menstrual cycle length is outside the normal range or bleeding is excessive: cycles outside 19 to 90 days are always abnormal, and most are 21 to 45 days even during the first postmenarcheal year. Continued menstrual abnormality in a hyperandrogenic adolescent for 1 year prognosticates at least 50% risk of persistence. Hyperandrogenism is best indicated by persistent elevation of serum testosterone above adult norms as determined in a reliable reference laboratory. Because hyperandrogenemia documentation can be problematic, moderate-severe hirsutism constitutes clinical evidence of hyperandrogenism. Moderate-severe inflammatory acne vulgaris unresponsive to topical treatment is an indication to test for hyperandrogenemia. Treatment of PCOS is symptom-directed. Cyclic estrogen-progestin oral contraceptives are ordinarily the preferred first-line medical treatment because they reliably improve both the menstrual abnormality and hyperandrogenism. First-line treatment of the comorbidities of obesity and insulin resistance is lifestyle modification with calorie restriction and increased exercise. Metformin in conjunction with behavior modification is indicated for glucose intolerance. Although persistence of hyperandrogenic anovulation for ≥2 years ensures the distinction of PCOS from physiologic anovulation, early workup is advisable to make a provisional diagnosis so that combined oral contraceptive treatment, which will mask diagnosis by suppressing hyperandrogenemia, is not unnecessarily delayed.

FIGURE 1

Comparison of the percent of menstrual cycles that are 21 to 45 days’ duration (red) and percent of menstrual cycles that are ovulatory (blue) by postmenarcheal age through young adulthood. Ovulation was determined by normalcy of urinary pregnanediol glucuronide in weekly samples collected during last 12 days of each menstrual cycle; cycles with clearly detectable but subnormal pregnanediol are designated here as having luteal insufficiency (green). It can be seen that most of the cycles that are not ovulatory had sufficient cyclic follicular activity to generate an immature corpus luteum, which indicates antecedent ovulation, rather than being truly anovulatory as the investigators had labeled them. Data from Metcalf et al.

FIGURE 2

Probability that an adolescent with symptomatic anovulatory symptoms will have ongoing menstrual abnormality. “All symptomatic anovulation” curve represents the data of Southam et al. “Hyperandrogenic” and “Nonhyperandrogenic” curves are hypothetical, based on data discussed in the text. Hyperandrogenic cases are predominantly a mix of physiologic anovulation and PCOS, with PCOS persisting. Nonhyperandrogenic cases are a mix of physiologic anovulation and hypogonadal cases, ranging from primary hypogonadism through hypothalamic amenorrhea to hypogonadotropic hypogonadism, with hypogonadal cases persisting.

FIGURE 3

Ferriman-Gallwey hirsutism scoring system. Each of the 9 body areas most sensitive to androgen is assigned a score from 0 (no hair) to 4 (frankly virile), and these separate scores are summed to provide a hormonal hirsutism score. Generalized hirsutism (score ≥8) is abnormal in the general US and UK populations, whereas locally excessive hair growth (score <8) is a common normal variant. The normal score is lower in Asian populations and higher in Mediterranean populations. Reproduced with permission from Martin KA, Chang RJ, Ehrmann DA, Ibanez L, Lobo RA, Rosenfield RL, et al. Evaluation and treatment of hirsutism in premenopausal women: An Endocrine Society Clinical Practice Guideline. J Clin Endocrin Metab. 2008;93:1105–1120.

FIGURE 4

Upper lip hirsutism scores(Ferriman-Gallwey) in adolescents and adults. Data in relation to menarcheal stage from Lucky et al; data in relation to age from Ferriman and Gallwey. Young adult FG scores are normally achieved 2 years after menarche.

FIGURE 5

Free testosterone plasma levels in normal postmenarcheal adolescent and adult female volunteers with normal ovarian morphology (V-NOM) compared with those with polycystic ovary morphology (V-PCOM) and PCOS. V-NOM and V-PCOM were healthy eumenorrheic females with no clinical signs of androgen excess. Data on these subjects were previously reported, but PCOM in adolescents has been redefined here as mean ovarian volume >12.0 mL, consistent with current consensus. Adolescents (Adol), 1 year postmenarcheal to 17.9 years of age, were similar to 18- to 39-year-old adults in each group. The free testosterone upper limit reference range (dotted line = 97th percentile = 9.3 pg/mL) was based on pooled adolescent and adult V-NOM, after excluding 1 outlier whose level was >3.0 SD from the mean of the entire group. V-PCOM had significantly higher free testosterone than pooled V-NOM (P = .03). Elevated levels were found in 2 of 6 adolescent and 4 of 30 adult volunteers with PCOM. To convert to pmol/L, multiply free testosterone by 3.47.

FIGURE 6

Ovarian androgenic function test results in normal postmenarcheal adolescent and adult female volunteers with normal ovarian morphology compared with those with polycystic ovary morphology and PCOS. Same groups as in Fig. 5. Adolescents (Adol) were similar to adults in each group. Dexamethasone 0.25 mg/m² orally was administered at 12 pm, and testosterone was measured 4 hours later (4 pm). This was followed shortly by administration of leuprolide acetate 10 μg/kg subcutaneously; 17-hydroxyprogesterone was sampled 20 to 24 hours later, 4 hours after a repeat 12 pm dexamethasone dose. Elevated total testosterone (>26 ng/dL) in response to a short dexamethasone androgen-suppression test (SDAST) or elevated 17-hydroxyprogesterone (>152 ng/dL) in response to a postdexamethasone gonadotropin-releasing hormone agonist (GnRHag) test indicate functional ovarian hyperandrogenism with 95% specificity and 68% concordance. Among the 93 PCOS patients, SDAST was abnormal in 85% (73% with abnormal GnRHag test), GnRHag test in 66% (92.5% with abnormal SDAST), and one or the other in 91%. Among volunteers with PCOM, 4 of 6 adolescents and 8 of 30 adults, including all with baseline elevation of free testosterone, had either an abnormal SDAST or GnRHag test result that is in the lower PCOS range. To convert to nanomole per liter, multiply total testosterone by 0.347 and 17-hydroxyprogesterone by 0.0303.