Highlights
In a multicenter Spanish registry of 269 patients with ACTH-dependent Cushing’s syndrome, urinary free cortisol expressed as times above the upper limit of normal (UFC×ULN) showed excellent discrimination between ectopic ACTH secretion (EAS) and Cushing’s disease (CD), with an area under the ROC curve of 0.90.
No patient with EAS had UFC×ULN below 3, supporting the practical value of mild cortisol excess as a strong argument for a pituitary source rather than ectopic ACTH production.
Severe biochemical hypercortisolism was much more typical of EAS: 40.5% of patients with UFC at least 10×ULN had ectopic disease, and combining UFC at least 10×ULN, late-night salivary cortisol at least 9×ULN, and hypokalemia identified 75% of EAS with 98% specificity.
The findings support a pragmatic diagnostic strategy: prioritize whole-body imaging when hypercortisolism is severe and hypokalemia is present, and favor pituitary-centered evaluation in milder cases.
Background
Differentiating EAS from CD remains one of the most difficult steps in the diagnostic pathway of ACTH-dependent Cushing’s syndrome. The distinction matters because the underlying diseases, imaging strategy, urgency of management, and operative approach differ substantially. CD typically arises from a corticotroph pituitary adenoma, whereas EAS is caused by nonpituitary ACTH-secreting tumors, often bronchial neuroendocrine tumors, pancreatic neuroendocrine neoplasms, thymic tumors, or small-cell lung cancer. EAS may present with rapidly progressive and severe hypercortisolism, but the phenotype can overlap with pituitary disease, especially when the ectopic source is occult or indolent.
Standard differential diagnosis often relies on a combination of clinical pattern recognition, pituitary MRI, high-dose dexamethasone suppression, CRH or desmopressin testing, and inferior petrosal sinus sampling (IPSS). Yet each of these tools has limitations. MRI may be negative or show incidentalomas. Dynamic tests lack perfect specificity. IPSS, while often considered the reference procedure for source localization in difficult cases, is invasive, technically demanding, and not universally available. In that context, a simple biochemical marker that helps estimate the probability of EAS before advanced testing would be clinically useful.
Urinary free cortisol is already central to diagnosing endogenous hypercortisolism. Expressing UFC as multiples of the local assay upper limit of normal offers a potentially transportable approach across laboratories, and prior work has suggested that very high UFC values favor ectopic disease. The Spanish validation study by Biagetti and colleagues addresses an important gap by testing this concept in an independent multicenter cohort and by examining whether UFC can be strengthened by late-night salivary cortisol and potassium status.
Study Design and Methods
Design and setting
This was a retrospective multicenter study using the Spanish Cushing Registry. Patients were diagnosed and managed in tertiary referral centers, an appropriate setting for a rare and complex disorder such as ACTH-dependent Cushing’s syndrome.
Participants
The analysis included 269 patients with ACTH-dependent Cushing’s syndrome: 208 with CD and 61 with EAS. This is a relatively large ectopic cohort for a validation study, reflecting the value of registry-based collaboration in rare endocrine disease.
Index markers and outcomes
The primary marker was UFC expressed as times above the upper limit of normal, or UFC×ULN. The investigators also assessed late-night salivary cortisol expressed in the same way, LNSC×ULN, and the presence of hypokalemia. The main outcome was the ability of these measures to discriminate EAS from CD, summarized by the area under the receiver operating characteristic curve and by clinically interpretable thresholds, including sensitivity, specificity, and predictive value.
Clinical framing
The study did not test UFC×ULN as a standalone replacement for established localization methods. Rather, it evaluated whether routine biochemistry can stratify pretest probability. That is an important distinction. In practice, the aim is not merely statistical classification but a more rational sequence of imaging and invasive procedures.
Key Findings
Patients with EAS had a distinctly more severe biochemical phenotype
Compared with patients with CD, those with EAS were older, with a median age of 59.0 years versus 44.9 years. More importantly, they showed far greater cortisol excess. Median UFC×ULN was 16.6 in EAS versus 3.6 in CD, and median LNSC×ULN was 9.3 versus 1.5. Both differences were highly statistically significant, with P values below 0.001.
These findings are biologically plausible. Ectopic ACTH-producing tumors often drive less regulated and more pronounced cortisol secretion, and they are more commonly associated with mineralocorticoid receptor activation by cortisol, which helps explain the link with hypokalemia.
UFC×ULN performed well as a discriminator
UFC×ULN achieved an area under the ROC curve of 0.90, which indicates excellent discrimination. LNSC×ULN performed similarly, with an AUC of 0.92. From a practical standpoint, this means that routine markers available early in the evaluation can already provide substantial information about disease source.
The most clinically striking threshold was the lower end of the UFC spectrum: no patient with EAS had UFC×ULN below 3. This is useful because it gives clinicians a reassuring boundary at the mild end of hypercortisolism. In a patient with ACTH-dependent Cushing’s syndrome whose UFC is elevated but remains below 3 times the upper limit of normal, ectopic disease appears very unlikely in this cohort.
At the opposite extreme, very severe cortisol excess increased the probability of ectopic disease substantially. Among patients with UFC at least 10×ULN, 40.5% had EAS. That does not make severe UFC elevation diagnostic of ectopic disease, because most such patients still had CD, but it is a major shift in pretest probability and should materially affect subsequent localization strategy.
Combining markers improved specificity for ectopic disease
The combination of UFC at least 10×ULN, LNSC at least 9×ULN, and hypokalemia identified 75% of EAS cases with 98% specificity. Clinically, this is the most actionable result in the paper. It suggests that when severe hypercortisolism is documented on both 24-hour urine and late-night saliva, and potassium is low, ectopic ACTH secretion should move to the front of the differential.
High specificity is particularly valuable here. Diagnostic delay in EAS can be dangerous because profound hypercortisolism drives infection risk, thromboembolism, muscle wasting, diabetes, neuropsychiatric complications, and cardiovascular instability. A highly specific biochemical triad may justify early whole-body imaging and aggressive cortisol-lowering therapy while the source is being confirmed.
Clinical Interpretation
How these findings can change the diagnostic sequence
The study supports a tiered, probability-based workup rather than a one-size-fits-all algorithm. In patients with relatively mild hypercortisolism, especially UFC below 3×ULN, the likelihood of EAS is very low based on this dataset. For such patients, pituitary MRI and pituitary-focused dynamic testing remain the most logical next steps, with IPSS reserved for unresolved cases.
By contrast, patients with marked UFC elevation, markedly abnormal late-night salivary cortisol, and hypokalemia should be treated as high risk for EAS. In these cases, early cross-sectional chest and abdominal imaging, and when appropriate functional imaging for neuroendocrine tumors, may be more efficient than an exclusively pituitary-first pathway. This is especially relevant when the clinical course is rapid or metabolically severe.
Why UFC×ULN is attractive in practice
UFC is already widely used in the diagnosis of Cushing’s syndrome. Expressing the result as a multiple of the assay-specific upper limit of normal partly addresses interassay variability and makes thresholds more portable across centers than absolute numeric values alone. It also creates a common language for clinicians: mild, moderate, and severe excess can be framed with immediate clinical meaning.
That said, UFC remains susceptible to collection errors, renal impairment, cyclic cortisol secretion, and assay-specific analytical issues. The value of UFC×ULN should therefore be interpreted in conjunction with method quality, repeat measurements, and the overall phenotype.
Strengths of the Study
The most important strength is external validation in an independent, real-world multicenter population. Diagnostic studies in rare endocrine disorders often rely on single-center derivation cohorts, which can overestimate performance. A registry-based national sample is therefore a meaningful advance.
The study also focuses on simple measurements that are already embedded in routine care. This increases translational value. Clinicians do not need new platforms or specialized biomarkers to apply the findings.
Finally, the combined analysis of UFC, LNSC, and hypokalemia reflects how endocrinologists actually reason at the bedside. Few localization decisions are made on one test in isolation, and the paper’s composite high-risk profile is likely to be more clinically durable than any single cutoff.
Limitations and Cautions
Several limitations should temper interpretation. First, the study was retrospective. That introduces the possibility of incomplete data, variability in timing of measurements, and center-level heterogeneity in diagnostic pathways.
Second, tertiary referral cohorts may enrich for more severe or atypical disease, potentially affecting threshold performance in general endocrine practice. Whether identical cutoffs will apply in community settings or in earlier-stage disease remains uncertain.
Third, while expressing hormone levels as multiples of the upper limit of normal improves comparability, it does not eliminate all laboratory differences. UFC and salivary cortisol assays vary in specificity, especially when immunoassays are used rather than mass spectrometry-based methods.
Fourth, the paper’s threshold-based interpretation should not be confused with absolute rules. EAS is heterogeneous. Some indolent neuroendocrine tumors can produce relatively modest cortisol excess, and some pituitary tumors can cause dramatic hypercortisolism. The study shows probabilities, not certainties.
Finally, the abstract does not provide full interval estimates for threshold-specific sensitivity and specificity. For clinical adoption, it will be important to review the complete manuscript for confidence intervals, calibration across centers, and details of reference standard confirmation.
Relationship to Current Guidance and Prior Literature
Current international guidance on the management of Cushing’s syndrome emphasizes a multimodal diagnostic approach and recognizes that the distinction between CD and EAS often requires integration of clinical, biochemical, radiologic, and invasive data. The Endocrine Society and Pituitary Society literature has repeatedly noted that hypokalemia and very severe hypercortisolism raise suspicion for ectopic ACTH secretion, but robust externally validated thresholds have been limited.
This Spanish study strengthens that long-held clinical impression with contemporary multicenter data. It does not displace IPSS or imaging, but it makes the earlier phase of workup more rational. In that sense, the paper fits well with modern diagnostic stewardship: use routine data to enrich or de-enrich the probability of a given source before moving to invasive or resource-intensive procedures.
Practical Takeaways for Clinicians
First, consider UFC×ULN as a quick probability tool rather than just a diagnostic confirmation marker for hypercortisolism. In ACTH-dependent disease, the magnitude of UFC elevation carries source information.
Second, UFC below 3×ULN appears strongly reassuring against EAS in this cohort. In such patients, a pituitary-focused pathway is usually reasonable unless other features strongly suggest ectopic disease.
Third, UFC at least 10×ULN should prompt heightened vigilance for EAS, particularly in older patients or those with rapid clinical deterioration.
Fourth, the combination of UFC at least 10×ULN, LNSC at least 9×ULN, and hypokalemia is highly specific for EAS and should accelerate whole-body imaging and consideration of urgent cortisol-lowering treatment.
Fifth, none of these thresholds should be applied without attention to assay method, collection quality, renal function, and the possibility of cyclic Cushing’s syndrome.
Conclusion
Biagetti and colleagues provide important external validation that the degree of cortisol excess, especially when expressed as UFC×ULN, helps distinguish EAS from CD in ACTH-dependent Cushing’s syndrome. The work is clinically useful because it converts an often qualitative impression into a practical, scalable framework. Mild UFC elevation favors a pituitary source; severe hypercortisolism, particularly when mirrored by markedly elevated late-night salivary cortisol and accompanied by hypokalemia, strongly points toward ectopic ACTH secretion.
For endocrinologists, internists, laboratory specialists, and multidisciplinary pituitary teams, the message is straightforward: routine biochemical severity should influence diagnostic sequencing. In high-risk patients, early whole-body imaging may be the more efficient path. In lower-risk patients, pituitary-centered evaluation remains appropriate. Future prospective studies should confirm how these thresholds perform across assay platforms and care settings and whether they reduce time to diagnosis, invasive testing, and cortisol-related morbidity.
Funding and ClinicalTrials.gov
The abstract provided does not report funding information or a ClinicalTrials.gov registration number. As a retrospective registry-based diagnostic study, formal trial registration may not have been applicable, but readers should consult the full article for complete disclosures.
References
1. Biagetti B, Marques P, Soto-Moreno A, García-Centeno R, González-Fernández L, Ollero García MD, Irigaray Echarri A, Cardona-Arias A, Moure Rodríguez MD, Paja M, Castro A, Manzano Valero L, Guerrero-Pérez F, Lamas C, Alcázar Lázaro V, Gracía P, Sanchis-Pascual D, Álvarez-Escolá C, Lozano-Aida C, Hanzu FA, Araujo-Castro M. Urinary Free Cortisol-Based Thresholds for Differentiating ACTH-Dependent Cushing’s: A Spanish Validation Study. The Journal of Clinical Endocrinology and Metabolism. 2026-04-30. PMID: 42063386.
2. Fleseriu M, Auchus R, Bancos I, Ben-Shlomo A, Bertherat J, Biermasz NR, Boguszewski CL, Bronstein MD, Buchfelder M, Carmichael JD, Castinetti F, Chanson P, Corcuff JB, Findling J, Gadelha M, Grossman A, Hanna F, Higham C, Ho K, Ioachimescu A, Kaiser U, Karavitaki N, Katznelson L, Kelly D, Lacroix A, McCormack A, Melmed S, Molitch M, Mortini P, Newell-Price J, Nieman L, Pereira AM, Petersenn S, Pivonello R, Raff H, Reincke M, Salvatori R, Scaroni C, Shimon I, Stratakis CA, Swearingen B, Varlamov EV, Witek P. Consensus on diagnosis and management of Cushing’s disease: a guideline update. Lancet Diabetes Endocrinol. 2021;9(12):847-875.
3. Nieman LK, Biller BMK, Findling JW, Murad MH, Newell-Price J, Savage MO, Tabarin A. Treatment of Cushing’s syndrome: an Endocrine Society Clinical Practice Guideline. J Clin Endocrinol Metab. 2015;100(8):2807-2831.
4. Newell-Price J, Grossman AB. Differential diagnosis of Cushing’s syndrome. Arq Bras Endocrinol Metabol. 2007;51(8):1199-1206.
