People with congenital adrenal hyperplasia (CAH) show extremely complex patterns of steroid profiles. Because of this, it’s increasingly clear that clinicians will be better able to diagnose and manage CAH by investigating multi-steroid profiles rather than by looking at a single analyte in the steroid pathway.

In our own work, we have found that liquid chromatography-tandem mass spectrometry (LC-MS/MS) is well-suited for creating such profiles relatively quickly and cheaply. In addition, incorporating unrecognized steroid analytes could provide improved disease control in patients with 21OHD, the most common form of CAH.

Clinical Case Part 1
A 26-year-old woman with classic 21-hydroxylase deficiency (21OHD) wants to have children. She has taken 3 doses of hydrocortisone plus fludrocortisone daily since childhood. This patient has regular monthly menses and no acne or recent adrenal crises. Despite regular unprotected intercourse for 6 months, she has not conceived. Laboratory testing shows a 17-hydroxyprogesterone (17OHP) of 5,000 ng/dL and a total testosterone (T) of 43 ng/dL. What additional laboratory testing is appropriate?

What CAH Patients Face

Individuals with CAH have severe problems with hormones produced by the adrenal glands. CAH is caused by autosomal-recessive genetic defects in cortisol biosynthesis. Loss of cortisol eliminates negative feedback on the hypothalamic-pituitary­ adrenal (HPA) axis, increasing adrenocorticotropin production. This results in hyperplasia of the adrenal cortex and dysregulated adrenal steroid production.

21OHD occurs as the classic (severe) form in 1:16,000 newborns and as the nonclassic (mild) form in 1:1,000 individuals. By definition, patients with classic 21OHD are cortisol deficient and—when there is no enzyme activity—also aldosterone deficient. Patients with nonclassic 21OHD have normal cortisol production, but at the expense of high precursor accumulation. In both forms, these precursors are diverted along various pathways, primarily to androgens. This means that all patients with 21OHD have some degree of androgen excess, with or without cortisol and aldosterone deficiency, depending on severity (Figure 1)

The Limitations of Traditional Tests

Laboratories traditionally have tested for 21OHD by measuring the cortisol intermediate 17-hydroxyprogesterone (17OHP)—the primary substrate for the 21-hydroxylase enzyme (P450 21A2). Values for serum 17OHP are typically <200 ng/dL in unaffected individuals and heterozygous carriers, 1,000–10,000 ng/dL in nonclassic 21OHD, and >10,000 ng/dL in classic 21OHD, with some variability.

Testing for 17OHP has its limits. For example, every state conducts newborn screening for 21OHD by measuring 17OHP via immunoassay (IA) on dried blood spots from Guthrie cards. Unfortunately, 17OHP also derives from the gonads, and several conditions, including prematurity, cause false-positive elevations of 17OHP. These analytical nuances aren’t confined to pediatrics. In adult women, ovarian production during the luteal phase of the menstrual cycle also causes 17OHP to rise. Ultimately, serum 17OHP is a sensitive, but not always specific, test for 21OHD.

Clinicians also screen for 17OHP to guide therapy in children with 21OHD, because 17OHP is a very sensitive measure of HPA axis control. But 17OHP has issues as a measure of disease control because of the marked and rapid fluctuations in values following a glucocorticoid dose. That’s especially true with short-acting hydrocortisone, the preferred therapeutic agent in these cases.

Progesterone (Prog) accumulates upstream of 17OHP not only in 21OHD but also in most other forms of CAH. While Prog is not used to diagnose 21OHD or to titrate therapy in children, elevated adrenal-derived Prog is an important cause of menstrual irregularity and infertility in women with 21OHD. When Prog accumulates in 21OHD, the adrenal enzyme 17-hydroxylase/17,20-lyase (P450 17A1) converts Prog not only to 17OHP but also about 30% to 16a-hydroxyprogesterone (16OHP), a side product that is not measured in most clinical laboratories. Similarly, the enzyme 11b-hydroxylase (P450 11B1) converts a small amount of Prog to 11b-hydroxy­progesterone (11OHP).

Aberrant metabolism to steroid­ side products also occurs with 17OHP. P450 11B1 readily 11b-hydroxylates 17OHP, which yields 21-deoxycortisol (21dF)—normally a minor adrenal product. Because P450 11B1 is highly expressed only in the adrenal cortex, 21dF is a metabolite that only comes from the adrenal glands. It’s a more specific marker of 21OHD than 17OHP, particularly for newborn screening. Commercial assays for 21dF have become available in reference laboratories in recent years, primarily for diagnosing classic and nonclassic 21OHD, but little data are available for its use in guiding therapy for the disease.

Now consider adrenal-derived androgens in 21OHD. Dehydroepiandrosterone sulfate (DHEAS) ordinarily is the single greatest steroid product of the adult adrenal cortex and the most abundant steroid in the circulation throughout life. With high androgen production, one might expect DHEAS to be high in nonclassic and even higher in classic 21OHD.

Patients with nonclassic 21OHD do have elevated serum DHEAS values as expected; however, patients with classic 21OHD paradoxically show normal or low DHEAS. In fact, DHEAS routinely falls below normal during therapy, even when all other biomarkers of control are elevated. In contrast, we recently showed that pregnenolone sulfate (PregS) is markedly elevated in classic 21OHD patients. The precursors that accumulate compete with the normal substrates not only for the steroid hydroxylases but also for the sulfotransferase (SULT2A1).

Ultimately, the pitfalls in measuring all these mono-, di-, and tri-hydroxypregnanes in patients with 21OHD are becoming clear. Many platform IAs are validated for serum samples containing normal physiologic distributions of active steroid hormones (such as cortisol) and their precursors and side-products­ (such as 17OHP and 21dF). In the serum of CAH patients, ratios of these steroids are dramatically distorted, and what would be trace metabolites in samples from healthy individuals become major components in the former.

Small percentage cross-reactivities­ now result in errors severe enough to alter CAH-related diagnoses. For example, the authors are aware of patients with classic 21OHD and 17-hydroxylase deficiency (17OHD) who have had serum cortisol values by IA reported in the 9–13 mg/dL range, consistent with nonclassic disease.

On a more positive note, LC-MS/MS assays subsequently confirmed the correct (classic) diagnoses with cortisol values <3 mg/dL. These assays are less prone to cross-reactivity­, but laboratories must identify and account for the minor metabolites even if their measurements are not reported. For example, 11-deoxycortisol, corticosterone, and 21dF are all isobaric dihydroxypregnanes, which are difficult to separate with standard reverse-phase high-performance LC systems. In serum from 21OHD patients, abundant 21dF can be miss-assigned as one of these other cortisol precursors if labs don’t pay meticulous attention to retention times, internal standards, and qualifier ions. We have found that biphenyl columns are superior to standard C8 or C18 columns for achieving the necessary resolution of these compounds.

Implementing LC-MS/MS Testing in Clinical Laboratories

At our institution, we have implemented LC-MS/MS assays for 17OHP, androstenedione (A4), testosterone (T), and DHEA.

Our older 17OHP and DHEA assays were manual IAs, and each analytical run took several hours to perform. This meant, for example, that performing a STAT 17OHP assay for a newborn with suspected CAH was challenging.

Our LC-MS/MS method is simple, cost-effective, and amenable to future automation. The sample preparation steps involve a protein precipitation followed by a solid phase extraction and analysis on an LC-MS/MS platform. In addition to superior specificity, the major advantage we’ve experienced from switching to LC-MS/MS has been efficiency. Our extraction process and LC method is common to 17OHP, A4, testosterone, and DHEA, enabling us to accurately determine all four analytes in a single analytical run.

This is an attractive approach, but comes with challenges. For example, A4 and T are only minimally separated on our LC-MS/MS method, and the ion pairs share a common product ion (289/109 for T and 287/109 for A4). Anyone running such an assay needs to pay careful attention to ensure the results from qualifying ion pairs are consistent with the quantifying ion pairs.

Additionally, validation studies need to investigate lack of interference from multiple steroids. Reference ranges may change when transitioning steroid assays from IA to LC-MS/MS formats, and large changes indicate a bias between the two methods. Clinicians absolutely should be notified of these references range changes so that they make sound clinical decisions based on results from two different assay methods.

Better Measures for Diagnosis and Monitoring

We’ve established that DHEAS is not a reliable biomarker of androgen excess in classic 21OHD, but what are the alternatives?

The conventional 19-carbon steroids A4 and T are normally minor products of the adrenal cortex, but the adrenal contribution varies with age and sex. In pre-pubertal children, aside from boys’ mini-puberty during the first 6 months of life, the adrenal glands are the major source of A4 and T. So elevated A4 and/or T in a child with 21OHD are reliable measures of poor disease control.

This is important because fastidious control of androgen production during childhood is essential to prevent premature development of secondary sexual characteristics, rapid somatic growth, and advanced skeletal maturation. In fact, estradiol (E2), derived from A4 and T, is the major stimulus for bone maturation, but the quantities of E2 required to advance bone age are below the limits of detection for even the best E2 assays today. Also, wide interlaboratory variations exist among various E2 IAs. For these reasons, 17OHP is used, in conjunction with A4, T, auxologic parameters, and bone age, as a surrogate for E2 in managing care for children with 21OHD.

The picture becomes more complicated in pubertal children and adults with 21OHD. In women, neither the adrenals nor the ovaries express an enzyme with good activity for the conversion of A4 to T, so the A4/T ratio is usually >1 regardless of source and absolute values. Elevated A4 and T signify either poor control of 21OHD or development of secondary polycystic ovary syndrome, caused by chronically high androgen exposure. The higher the A4/T ratio and the higher the A4 concentration, the more likely the androgens are derived from the adrenal cortex.

For example, a woman with classic 21OHD, an A4 of 120 ng/dL and a T of 80 ng/dL is likely to have ovarian androgen excess. In contrast, an A4 of 400 ng/dL and T of 65 ng/dL indicate poor control of 21OHD. Because 98–99% of T is protein bound in the circulation, it is important to assess the free and bio-available fractions of T as well to gauge the biological potency of a given T value. Measurement of sex hormone-binding globulin in conjunction with a total T measurement by LC-MS/MS allows reliable calculation of free and bioavailable T.

Men with 21OHD face similar complications. A normal serum T value alone can be misleading, as both normal testes and the adrenal cortex in poorly controlled 21OHD produce large amounts of T. Because only the Leydig cell of the testes expresses 17b-hydroxysteroid dehydrogenase type 3—the only human enzyme that efficiently converts A4 to T—the A4/T ratio in men is <1, typically <0.4. In practical terms, when a man’s serum A4 concentration exceeds that of T, it’s a sign of poor disease control, and androgens are primarily derived from the adrenal cortex.

To further assess the HPA and testicular axes, serum luteinizing hormone will be suppressed when adrenal-derived androgens are high. A semen analysis, the ultimate assessment of testicular function, is often abnormal in men with 21OHD. Testicular adrenal rest tumors develop in 30–50% of men with classic 21OHD, and these tumors and/or elevated follicle-stimulating hormone are both negative prognostic factors for fertility in men with 21OHD.

This situation is similar to the metabolism of 17OHP to 21dF in the adrenal cortex of patients with 21OHD, because androgens are substrates for P450 11B1. In fact, the three most abundant steroids identified in samples of adrenal vein blood are DHEAS, cortisol, and 11b-hydroxyandrostenedione (11OHA4), which is the major P450 11B1 metabolite of A4 substrate.

Once again, LC-MS/MS proves useful. Using this technology, we profiled four 11-oxygenated, 19-carbon steroids—11OHA4, 11b-hydroxyT, 11-ketoA4, and 11-ketoT (collectively 11oxC19 steroids)—in serum from patients with 21OHD and controls. We found that serum concentrations of 11OHA4 and 11-ketoT were three- to four-fold higher in patients with 21OHD than in matched controls and averaged twice as high as the patients’ A4 and T levels. In women with 21OHD, all 11oxC19 steroids correlate with A4 and T, and all serve as biomarkers of poor disease control. In men with 21OHD, 11-ketoT correlates inversely with T, reflecting the adrenal as the sole origin of 11-ketoT and the utility of this biomarker in determining the source of T in men with 21OHD.

Bearing all this in mind, we believe 11oxC19 steroids will be valuable measures for diagnosing and monitoring disease in patients with 21OHD. Furthermore, 11-ketoT—not T—is the major circulating androgen in children, women, and men with classic but poorly controlled 21OHD.

Steroid Panels Show Promise

The preceding discussion suggests that steroid panels might be more useful than single analytes in diagnosing classic and nonclassic 21OHD and in managing these patients into adulthood. Reference laboratories now offer steroid panels using LC-MS/MS methods, which enables accurate measurement of multiple steroids from a single small serum sample.

Some of the challenges in developing these panels include balancing speed with adequate separation, obtaining necessary isotopically labeled internal standards, sacrificing some sensitivity for overlapping analytes, and maintaining quality in the face of limited demand.

Despite these challenges, the advantages to patients, physicians, and investigators are considerable, and improvements in instrumentation and data analysis will render these panels more practical. Nearly all steroid assays in reference laboratories today employ LC-MS/MS.

Clinical Case Part 2:
In a woman with classic 21OHD who is attempting to become pregnant, follicular-phase Prog is the critical analyte to optimize fertility. A4 testing will help explain the source of T and help guide her treatment. Not mentioned in Part 1, electrolytes and plasma renin activity or mass are used to titrate mineralocorticoid replacement with fludrocortisone acetate. Table 2 provides a list of typical current laboratory tests for patients with 21OHD.

Urinary Steroid Metabolites and Saliva Steroids

Long before reference laboratories embraced LC-MS/MS for serum steroid measurements, investigators used gas chromatography-MS to profile urine steroid metabolites. This procedure is highly informative but also very labor-intensive. It requires deconjugation of the steroid sulfates and glucuronides and derivatization prior to analysis.

Steroid measurements from a 24-hour urine collection have the advantage of reflecting an integrated measure of control throughout the day, but patients—especially children—have difficulty complying with 24-hour urine collections. The major metabolites of 17OHP, 21dF, and A4 plus T—pregnanetriol, pregnanetriolone, and androsterone plus etiocholanolone, respectively—are the major urinary metabolites relevant for 21OHD. Saliva measurements of cortisol and 17OHP are now routinely available, and their potential use in disease monitoring requires further exploration.

Other Forms of CAH

The concept of disordered steroidogenesis in other forms of CAH is the same as in 21OHD: high steroids above the block in the steroid pathway, low steroids below the block, and increased flux around the block. In 17OHD, a block orthogonal to that in 21OHD causes mineralocorticoid excess and loss of androgen (and estrogen) synthesis. In 11b-hydroxylase deficiency (11OHD), a block in the last step of cortisol production causes widespread accumulation and overproduction of all steroids but cortisol, resulting in both androgen and mineralocorticoid excess. These conditions are summarized in Table 1. 

New Biomarkers and the Future

CAH patients—depending on CAH type, age, sex, and disease management—show unique and extremely complex patterns of steroid profiles. Laboratory diagnosis and management of CAH relies on investigation of multi-steroid profiles rather than single components in the steroid pathway.

LC-MS/MS is particularly suited to produce such profiles in a cost- and time-efficient manner. Additionally, incorporating unrecognized steroid analytes such as PregS, 21dF, 16OHP, 11OHA4, and 11-ketoT into steroid panels along with Prog, 17OHP, A4, and T holds the potential for more reliable assessment of short-term and long-term disease control in patients with 21OHD. Such panels may afford a diagnosis of nonclassic 21OHD without cosyntropin stimulation testing, too.

As LC-MS/MS and steroid panels enter the mainstream, diagnosing and managing CAH will become more direct and precise.

Suggested Reading

1.       Auchus RJ. The classic and nonclassic congenital adrenal hyperplasias. Endocr Pract 2015;21:383-89.

2.       Auchus RJ, Arlt W. Approach to the patient: the adult with congenital adrenal hyperplasia. J Clin Endocrinol Metab 2013;98:2645-55.

3.       Casteràs A, De Silva P, Rumsby G, et al. Reassessing fecundity in women with classical congenital adrenal hyperplasia (CAH): normal pregnancy rate but reduced fertility rate. Clin Endocrinol (Oxf) 2009;70:833-37.

4.       Charmandari E, Matthews DR, Johnston A, et al. Serum cortisol and 17-hydroxyprogesterone interrelation in classic 21-hydroxylase deficiency: is current replacement therapy satisfactory? J Clin Endocrinol Metab 2001;86:4679-85.

5.       King TF, Lee MC, Williamson EE, et al. Experience in optimizing fertility outcomes in men with congenital adrenal hyperplasia due to 21 hydroxylase deficiency. Clin Endocrinol (Oxf) 2016;84:830-36.

6.       Krone N, Hughes BA, Lavery GG, et al. Gas chromatography/mass spectrometry (GC/MS) remains a pre-eminent discovery tool in clinical steroid investigations even in the era of fast liquid chromatography tandem mass spectrometry (LC/MS/MS). J Steroid Biochem Mol Biol 2010;121:496-504.

7.       Minutti CZ, Lacey JM, Magera MJ, et al. Steroid profiling by tandem mass spectrometry improves the positive predictive value of newborn screening for congenital adrenal hyperplasia. J Clin Endocrinol Metab 2004;89:3687-93.

8.       Nakamura Y, Hornsby PJ, Casson P, et al. Type 5 17β-hydroxysteroid dehydrogenase (AKR1C3) contributes to testosterone production in the adrenal reticularis. J Clin Endocrinol Metab 2009;94:2192-98.

9.       Turcu AF, Nanba AT, Chomic R, et al. Adrenal-derived 11-oxygenated 19-carbon steroids are the dominant androgens in classic 21-hydroxylase­ deficiency. Eur J Endocrinol 2016;174:601-09.

10.     Turcu AF, Rege J, Chomic R, et al. Profiles of 21-carbon steroids in 21-hydroxylase deficiency. J Clin Endocrinol Metab 2015;100:2283-90.

Adina Turcu, MD is an assistant professor of internal medicine at Michigan Medicine, University of Michigan in Ann Arbor, Michigan. +Email: [email protected]

Hema Ketha, PhD is a clinical instructor in the department of clinical chemistry and toxicology, director of toxicology and drug analysis, and associate director of clinical chemistry at Michigan Medicine, University of Michigan in Ann Arbor, Michigan. +Email: [email protected]

Richard Auchus, MD, PhD is a professor of internal medicine and pharmacology  and director of the division of metabolism, endocrinology, and diabetes fellowship program at Michigan Medicine, University of Michigan in Ann Arbor, Michigan. +Email: [email protected]