LC-MS/MS Redefines Accuracy in Congenital Adrenal Hyperplasia Newborn Screening: Reducing the False-Positive Burden

Highlights

1. Liquid chromatography-tandem mass spectrometry (LC-MS/MS) significantly outperforms radioimmunoassay (RIA) in serum confirmatory tests for Congenital Adrenal Hyperplasia (CAH), yielding a positive predictive value (PPV) of 49% compared to 30%.

2. The (17OHP+Δ4)/cortisol ratio emerged as the most discriminative marker, achieving a 100% PPV at a cutoff of 24.9, effectively eliminating false-positive results in this cohort.

3. Serum 21-deoxycortisol (21DF) measured by LC-MS/MS showed high specificity (99%), serving as a robust diagnostic alternative to 17-hydroxyprogesterone (17OHP).

4. Genetic testing of the CYP21A2 gene remains a critical tool for resolving indeterminate biochemical results, particularly in asymptomatic cases with persistently elevated steroid levels.

Background: The Challenge of False-Positives in CAH Screening

Congenital Adrenal Hyperplasia (CAH), primarily caused by 21-hydroxylase deficiency, is a group of autosomal recessive disorders that impair cortisol and aldosterone synthesis. The resulting accumulation of 17-hydroxyprogesterone (17OHP) and adrenal androgens can lead to life-threatening salt-wasting crises and ambiguous genitalia in newborns. To prevent these outcomes, newborn screening (NBS) programs worldwide utilize 17OHP measurements from dried blood spots (DBS).

However, CAH-NBS is notorious for its high rate of false-positive (FP) results. In the neonatal period, cross-reactivity with fetal zone steroids and the physiological stress of premature or ill infants can lead to transiently elevated 17OHP levels. These false positives cause significant parental anxiety, increase healthcare costs, and place an undue burden on pediatric endocrine clinics. Traditionally, radioimmunoassay (RIA) has been the workhorse for serum confirmation, but its susceptibility to interference has limited its efficacy. The emergence of liquid chromatography-tandem mass spectrometry (LC-MS/MS) offers a more specific alternative, though large-scale comparative data in prospective cohorts have been limited until recently.

Study Design and Methodology

In a landmark prospective longitudinal cohort study, Carvalho et al. (2026) evaluated the efficacy of confirmatory testing methodologies across a massive population of 708,437 newborns. The primary objective was to compare the diagnostic performance of RIA and LC-MS/MS in reducing the ‘drop-off’ of false positives in the screening pipeline.

The screening protocol followed a two-tier approach. Newborns with initial neonatal-17OHP (N17OHP) levels ≥ 2 times the 99.5th percentile were recalled for serum confirmatory testing. These recalled infants underwent parallel testing using two distinct methodologies:

1. Radioimmunoassay (RIA)

This method measured serum 17OHP levels, a traditional standard in many clinical laboratories.

2. Liquid Chromatography-Tandem Mass Spectrometry (LC-MS/MS)

This advanced platform allowed for the simultaneous quantification of multiple steroid metabolites, including 17OHP, 21-deoxycortisol (21DF), androstenedione (Δ4), and cortisol.

The researchers calculated positive predictive values (PPV), used non-parametric tests for group comparisons, and constructed Receiver Operating Characteristic (ROC) curves to determine optimal diagnostic cutoffs. Infants with persistently altered hormonal results were subjected to CYP21A2 genotyping to confirm the diagnosis of classical or non-classical CAH.

Key Findings: RIA vs. LC-MS/MS Performance

The study reported a recall rate of 0.03% for the first-tier N17OHP screening. Among the recalled newborns, serum 17OHP levels remained elevated in 26% of those tested via RIA, but only in 11% of those tested via LC-MS/MS. This discrepancy highlights the overestimation inherent in RIA due to its lower specificity and cross-reactivity with other steroids.

The diagnostic performance metrics were starkly different between the two methods:

• The PPV for RIA was 30%.

• The PPV for LC-MS/MS was 49%.

Out of the total cohort, 58 newborns were diagnosed with classical forms of CAH. Interestingly, 32 asymptomatic newborns exhibited persistently increased serum 17OHP levels, highlighting the ‘gray zone’ often encountered in screening programs where biochemical markers do not immediately correlate with clinical symptoms.

Refining Diagnostic Accuracy: Steroid Ratios and Cutoffs

The most significant contribution of this research lies in the identification of highly specific cutoffs and steroid ratios. The ROC-curve analysis demonstrated that 17OHP measured by LC-MS/MS had the highest diagnostic performance for a single marker. A cutoff of 48.3 ng/mL (146.3 nmol/L) yielded 100% sensitivity, ensuring that no true cases were missed.

However, single-marker measurements were surpassed by the use of steroid ratios. The ratio of (17OHP + Δ4) to cortisol proved to be the superior diagnostic tool. By using a cutoff of 24.9, the researchers achieved a 100% PPV. This ratio is biologically plausible; it reflects the accumulation of precursors (17OHP and androstenedione) relative to the deficiency of the end-product (cortisol), effectively filtering out infants who may have high 17OHP due to stress rather than enzyme deficiency.

Additionally, serum 21-deoxycortisol (21DF) was confirmed as a highly specific marker for 21-hydroxylase deficiency. Since 21DF is only produced when 17OHP is diverted into the pathway due to a block at the 21-hydroxylase step, its presence is nearly pathognomonic for CAH. The study found its specificity (99%) to be comparable to 17OHP measured by LC-MS/MS.

Expert Commentary: Clinical Implications

The findings by Carvalho et al. underscore a necessary shift in the management of CAH newborn screening. While immunoassays are cost-effective for the initial high-volume screening of dried blood spots, they are insufficient for serum confirmation. The transition to LC-MS/MS for second-tier testing or serum confirmation should be considered the clinical gold standard.

From a health policy perspective, the reduction of false positives from 70% (using RIA) to approximately 50% (using LC-MS/MS) is a significant improvement, but the 100% PPV achieved by the (17OHP+Δ4)/cortisol ratio is the real breakthrough. Implementing this ratio in clinical algorithms could virtually eliminate the ‘false alarm’ scenarios that currently plague NBS programs.

One limitation noted in the study is the persistent elevation of 17OHP in some asymptomatic newborns. In these cases, the researchers emphasize that genotyping is not just a secondary tool but a primary necessity to resolve indeterminate biochemical results. Molecular analysis of the CYP21A2 gene can distinguish between carriers, non-classical CAH, and true false-positives that might be caused by rare variants or maternal factors.

Conclusion: A New Standard for Confirmatory Testing

In conclusion, the measurement of serum 17OHP by LC-MS/MS offers superior diagnostic performance for CAH confirmatory testing compared to RIA. The integration of steroid ratios—specifically (17OHP+Δ4)/cortisol—provides a robust mechanism to maximize PPV and minimize the clinical and psychological fallout of false-positive results. For clinicians, the message is clear: when faced with a positive screening result, LC-MS/MS-based confirmation and the calculation of steroid ratios should be the immediate next step, followed by genotyping for any cases that remain biochemically ambiguous.

References

1. Carvalho DF, Lima-Valassi HP, Hayashi GY, et al. Efficacy of liquid-chromatography and radioimmunoassay in false-positives’ drop-off in CAH newborn screening. The Journal of clinical endocrinology and metabolism. 2026. PMID: 41811754.

2. Speiser PW, Arlt W, Auchus RJ, et al. Congenital Adrenal Hyperplasia Due to Steroid 21-Hydroxylase Deficiency: An Endocrine Society Clinical Practice Guideline. J Clin Endocrinol Metab. 2018;103(11):4043-4088.

3. Gidlöf S, Falhammar H, Thilén A, et al. One hundred years of congenital adrenal hyperplasia: From founder mutations to newborn screening. J Intern Med. 2013;274(2):118-143.