Consensus on the Diagnosis of Cushing’s Disease: A Collaborative Statement from the Korean Endocrine Society and Japan Endocrine Society

1. Clinical context

In both Korea and Japan, serum cortisol is predominantly measured using automated immunoassay platforms, although the specific analyzers and reagents may vary among institutions, leading to modest inter-assay variability. However, these minor methodological differences do not sufficiently explain the discrepancies in cutoff values between the two national guidelines, which primarily reflect distinct diagnostic strategies rather than assay methodology. It should also be noted that mass spectrometry- based cortisol measurements have not yet been widely adopted in either country. Cortisol secretion follows a pulsatile pattern that is regulated by the circadian rhythm of ACTH. In healthy individuals, this diurnal variation is well established, with peak cortisol levels in the early morning and the lowest levels around midnight [14]. The measurement of basal ACTH concentrations is particularly useful in distinguishing between pituitary-dependent (CD) and ACTH-independent CS [15]. Because plasma ACTH is highly unstable, appropriate pre-analytical handling, including use of pre-chilled tubes, immediate placement on ice, and prompt centrifugation, is essential to prevent degradation and avoid falsely low values that may lead to diagnostic misclassification. In clinical practice, morning ACTH levels <10 pg/mL (2 pmol/L) with normal or high cortisol levels strongly suggest ACTH-independent CS, and ACTH levels >20 pg/mL (4 pmol/L) indicate ACTH-dependent CS [16]. However, the KES guidelines do not explicitly include morning ACTH measurements as a required screening test for CS. Instead, the KES emphasizes biochemical suppression tests and diurnal cortisol variation for initial hypercortisolism detection. A Korean study found that plasma ACTH levels have significant overlap between CD and adrenal CS, which limits their diagnostic utility [17]. That study reported that although low ACTH levels traditionally indicate adrenal CS, a considerable proportion of adrenal CS patients still had detectable ACTH levels (3.5 to 15.6 pmol/L), making differentiation unreliable. Another study conducted at a tertiary center in Korea analyzed baseline ACTH levels and diagnostic markers in patients with CD [18]. That study found that preoperative ACTH levels did not differ significantly between macro-tumors (101.5±23.2 pg/mL) and microtumors (83.6±11.1 pg/mL, P=0.44), demonstrating a lack of correlation between ACTH levels and tumor size. Additionally, the ACTH-to-cortisol ratio did not predict remission after transsphenoidal surgery, further limiting its clinical utility in decision-making. Given those limitations, the KES guidelines deemphasize morning ACTH levels as a core screening parameter due to inconsistent results between CD and adrenal CS. In contrast, the JES guidelines pose early morning plasma ACTH and serum cortisol levels as reference indicators for proceeding to other screening tests. Specifically, the presence of normal to high ACTH and cortisol levels in the context of clinical features suggestive of ACTH-dependent CS supports proceeding with further diagnostic evaluation. In contrast, clearly low levels of both ACTH and cortisol suggest that endogenous CS is unlikely, and in such cases, an additional workup is generally not recommended. It should be noted, however, that the guidelines do not define a precise lower limit for ACTH and cortisol, and further studies are needed to clarify those thresholds.

1. Clinical context

In physiological conditions, serum cortisol levels exhibit a circadian rhythm, reaching a nadir around midnight. Measuring cortisol levels during the late-night nadir provides information that is valuable for diagnosing CS and differentiating it from other causes of hypercortisolism.

1. Clinical context

UFC has classically been the CD screening test most used in Korea. In contrast, UFC is regarded as a supportive laboratory finding for the diagnosis in Japan. It is of note that the method for measuring UFC in Japan changed in 2022 from a radioimmunoassay (RIA) to a chemiluminescent immunoassay (CLIA), which is known to yield lower values than the RIA. Because the cortisol excreted in urine is collected for a full 24 hours, the UFC avoids the confounding effects of diurnal cortisol fluctuations and is not affected by changes in CBG levels [19,20]. A positive UFC result is typically defined as >90 μg/day when measured by RIA or >50 μg/day using high-performance liquid chromatography or CLIA in at least 2–3 repeated tests and concurrent 24-hour urinary creatinine/volume measurement to ensure accuracy [10]. On the other hand, UFC testing has several limitations. UFC values can fluctuate due to renal function, hydration status, and inter-assay variability [19,21]. Additionally, in cases of mild or cyclic CD, cortisol secretion can intermittently fall within normal ranges, leading to potential false-negative results [22]. Obesity, depression, and alcoholism can all cause mild elevations in UFC, leading to false positives [23]. The Cowork of Adrenal Research project in Korea, a large prospective multicenter study, found that UFC levels showed notably poor diagnostic performance. That study highlighted that UFC might not reliably detect subtle cortisol alterations and those technical challenges, such as assay variability and collection inconsistencies, can limit its accuracy [24]. Interestingly, emerging data suggest that Asian patients, including Koreans, can present with inherently higher baseline UFC levels and more severe disease manifestations than non-Asians. A pooled analysis indicated that 39.3% of Asian patients had severe disease (UFC >5× upper limit of normal) versus 20.1% in non-Asians [25]. This disparity can be attributed to a higher prevalence of macro-tumors, which lead to increased cortisol production in Asian populations [26–28]. Given those variations, ethnicity-specific UFC thresholds and further validation of cutoff values in Asian cohorts are needed to improve diagnostic accuracy.

1. Clinical context

LNSC measurement is a non-invasive and reliable method for assessing free cortisol levels, which are in equilibrium with the unbound, biologically active form of cortisol in serum [29,30]. This equilibrium exists because free cortisol diffuses through the acinar cells of the salivary glands into the oral cavity, with only a small fraction (approximately 15%) bound to CBG [31]. Because LNSC reflects only the free fraction, its levels are unaffected by binding proteins and accurately follow the physiological circadian rhythm, including the early morning peak and late-night nadir [32]. Moreover, saliva sampling is simple and non-invasive, which minimizes stress-induced cortisol elevations. LNSC is considered to be a practical and patient-friendly tool for repeated assessment and longitudinal monitoring of hypothalamic–pituitary–adrenal axis function. Despite international acceptance of LNSC as a screening tool, its clinical application remains limited in Korea, partly due to the lack of standardized cutoff values. A systemic literature review reported that the optimal cutoff values for LNSC vary widely (range 0.13 μg/dL [3.6 nmol/L] to 0.55 μg/dL [15.2 nmol/L]), with a pooled sensitivity of 92% (95% confidence interval [CI], 88% to 94%) and specificity of 96% (95% CI, 94% to 97%) [33]. A study from Taiwan showed that a cutoff for LNSC level of 0.17 μg/dL (4.69 nmol/L) was associated with a sensitivity of 98% and specificity of 100% [34]. In a Japanese study, LNSC was measured using three different assays—RIA, electro-chemiluminescence, and enzyme-linked immunosorbent assay—across four facilities. With an absolute cutoff value of 0.4 μg/dL (11.04 nmol/L), the sensitivity and specificity were 86% and 100%, respectively. To address variability across assays and enable standardization, those investigators evaluated a ratio-based cutoff set at 1.5 times the mean value in healthy controls, and it yielded a sensitivity of 96% and specificity of 88%. Those findings highlight the need for assay standardization in the interpretation of LNSC measurements [35]. It is important to note that LNSC is not recommended for individuals with an altered circadian rhythm, such as night-shift workers [9]. The KES guidelines currently do not specify a definitive LNSC cutoff, reflecting the ongoing need for large-scale validation studies. However, efforts to establish reference values and promote the clinical application of LNSC testing in Korea are ongoing. Although the JES diagnostic guidelines acknowledge the usefulness of LNSC in Japan, the measurement method has not been standardized, and reference values have yet to be established. In addition, salivary cortisol testing is not currently covered by insurance.

1. Clinical context

The 1-mg ODST is used most widely when screening for CS [36]. In this test, 1-mg of dexamethasone is administered orally at 11 PM, and serum cortisol levels are measured between 8:00 AM and 9:00 AM the following morning. A serum cortisol level >1.8 μg/dL (50 nmol/L) suggests the failure of suppression, indicating possible hypercortisolism [8,9,37]. The KES guidelines align with global recommendations, adopting the 1.8 μg/dL cutoff due to its high sensitivity (~95%). This threshold was established based on studies demonstrating that post-dexamethasone cortisol levels above 1.8 μg/dL are associated with increased morbidity and mortality, even in patients without overt CS [38]. Additionally, using a higher cutoff (5 μg/dL) risks underestimating the clinical significance of mild autonomous cortisol secretion, which has been linked to cardiovascular disease, osteoporosis, and metabolic abnormalities [39]. In Korea, most clinical studies evaluating the ODST have also used the 1.8 μg/dL cutoff as the primary reference value, reflecting its established diagnostic accuracy and clinical applicability [18,24,27,40]. Despite its high sensitivity, the ODST has limitations due to the potential for false positives in patients with obesity, alcoholism, or psychiatric disorders [16]. Therefore, it is recommended that the ODST be interpreted alongside additional screening tests to improve diagnostic specificity.

In Japan, a dose of 0.5-mg of dexamethasone has been used for the ODST. Japanese prior data indicate that a 0.5-mg ODST is more sensitive than the 1-mg ODST in the Japanese. This higher sensitivity has been attributed to population-specific factors such as leaner body habitus and greater cortisol suppressibility in response to dexamethasone, leading to adoption of the 0.5 mg dose with a higher cutoff for screening [41]. A cortisol cutoff value of 5 μg/dL for diagnosing ACTH-dependent CS showed a sensitivity of 94.2% with a 1-mg dose and 99.1% with a 0.5-mg dose [42]. Furthermore, to enable the early detection of mild hypercortisolism, the JES uses a cutoff value of 3 μg/dL for screening subclinical CD. Thus, the 0.5‑mg ODST is characterized by higher sensitivity for mild or subclinical disease, at the expense of a potential reduction in specificity and an increased need for confirmatory testing.

Overall, the differing preferences of the KES and JES reflect variations in historical validation studies, population-specific responses to dexamethasone, and diagnostic priorities rather than differences in assay methodology. The 1-mg ODST offers broad international comparability and robust sensitivity, whereas the 0.5-mg ODST provides increased sensitivity for mild disease at the potential cost of specificity, necessitating careful clinical correlation.

1. Clinical context

The desmopressin test is emerging as a valuable tool in the diagnostic evaluation of CS, particularly in distinguishing CD from non-neoplastic hypercortisolism (NNH) [43] and ectopic ACTH syndrome (EAS). Desmopressin acts on arginine vasopressin (AVP) receptor 1b, which is highly expressed in corticotrophic tumors, or ectopically expressed AVP receptor 2, which stimulates ACTH and cortisol release, making it a useful functional test in the differential diagnosis of ACTH-dependent CS [44,45]. A positive response (ACTH increase >50% and cortisol increase >20%) strongly suggests CD, and a lack of response suggests NNH or EAS. A recent systematic review and meta-analysis, which included an Asian cohort, assessed the accuracy of intravenous desmopressin testing in differentiating CD from NNH and EAS. The combined criteria of ΔACTH >35% and Δ cortisol >20% showed a sensitivity of 85% and specificity of 64% [46]. Beyond diagnosis, the desmopressin stimulation test is increasingly used for postoperative monitoring in CD patients [47,48], and the KES guidelines note that an exaggerated ACTH and cortisol response to desmopressin post-surgery can indicate residual tumor tissue and predict disease recurrence. This approach aligns with international trends, in which desmopressin used as an alternative to the CRH test, which is becoming less available globally [49]. The desmopressin test protocols of both the KES and JES use a 4-μg dose of desmopressin, in contrast to the 10-μg dose commonly used in Western countries. This lower-dose protocol in East Asia is based on early Japanese clinical studies demonstrating that 4 to 5 μg produced a clear stimulatory response in most patients with CD [41]. These data were subsequently incorporated into the early JES recommendations and later into the 2015 KES guideline, which specifies a 4-μg dose. Although no head-to-head comparison between 4–5 μg and 10 μg dosing has been conducted, accumulated regional experience indicates that the lower dose provides adequate diagnostic stimulation while minimizing vasopressin-related adverse effects. In contrast, the 10-μg dose became standard in Western centers largely because early desmopressin stimulation protocols adopted this dose to maximize ACTH responsiveness [50].

1. Clinical context

The CRH test is a dynamic function test that assesses the pituitary response to CRH [46]. In most cases of CD, CRH stimulates ACTH secretion, leading to a subsequent rise in cortisol levels. This response is less pronounced in EAS, making the test useful in differentiating CD from other ACTH-dependent causes of hypercortisolism [51]. The KES and JES both incorporate the CRH test into their diagnostic frameworks. The test involves the intravenous administration of 100-μg of CRH after an overnight fast, followed by the measurement of plasma ACTH and serum cortisol levels at 15-, 30-, and 60-minutes post-injection [8,9,52]. A ≥50% increase in ACTH and ≥20% increase in cortisol within 30 minutes of CRH administration is suggestive of CD.

The differential response to CRH between CD and EAS reflects their underlying pathophysiology. Corticotroph tumors in CD overexpress CRH receptor 1 (CRHR1) and retain sensitivity to hypothalamic stimulation, leading to a marked rise in ACTH and cortisol after CRH administration [53]. In contrast, most EAS demonstrate minimal or absent CRHR1 expression and synthesize ACTH autonomously, independent of hypothalamic regulation [54]. Because EAS do not rely on CRH-mediated signaling, they exhibit little or no biochemical response to CRH administration. These differences in CRH-receptor expression and regulatory dependence on the hypothalamic–pituitary axis form the principal mechanistic basis for the diagnostic utility of the CRH stimulation test in distinguishing CD from EAS. Furthermore, this test is contingent on the presence of hypercortisolism, a state in which normal corticotrophs no longer synthesize ACTH in a CRH-dependent manner.

However, false-negative results can occur in cases of macro-tumors, and false-positive results can be seen in EAS, particularly in bronchial carcinoids. Therefore, caution is warranted, and clinical decisions should not be based on this test alone. In Japan, the CRH test is a routine diagnostic tool, but in Korea, it is rarely performed due to cost constraints.

1. Clinical context

Currently, both the low-dose dexamethasone suppression test (LDDST) and HDDST are still widely used for the diagnosis and differential diagnosis of CS. The LDDST typically involves oral administration of 0.5-mg of dexamethasone every 6 hours for 2 days, followed by the measurement of serum cortisol [1]. However, its diagnostic performance of is comparable to that of the 1-mg ODST, so its clinical utility is gradually declining [55]. The HDDST has been a traditional tool for distinguishing CD from adrenal CS or EAS. It requires administration of dexamethasone over 2 days, commonly 2 mg every 6 hours for 2 consecutive days (total 16 mg). Serum cortisol and 24-UFC is measured before and after dexamethasone administration. Compared with the HDDST, the 8-mg ODST is more convenient due to its single-dose administration and lack of requirement for 24- hour urine collection [56]. The 8-mg ODST has been reported to have a sensitivity of approximately 95%, making it more effective than the traditional HDDST in distinguishing between CD and EAS [57,58]. Although the 8-mg ODST is increasingly being adopted as a convenient and efficient alternative for diagnosing CD, the HDDST continues to be commonly used in clinical practice in Korea. The KES recommends that the HDDST be interpreted based on an absolute post-suppression cortisol level <5 μg/dL, which is considered indicative of CD. The JES recommends >50% suppression from the baseline cortisol level after the 8-mg ODST as the diagnostic threshold for CD.

Given current differences in diagnostic practice and population characteristics, additional studies in Asian populations are warranted to validate the diagnostic performance of the 8-mg ODST and further clarify its role relative to the HDDST. Caution is required in cases of ectopic ACTH-secreting pheochromocytoma because high-dose dexamethasone can trigger a hypertensive crisis in those patients.

1. Clinical context

The KES guideline’s recommendation to used sella MRI as the primary imaging modality for diagnosing CD is well-supported by both international and Korean studies. Sella MRI provides high anatomical resolution for detecting pituitary tumors and assessing structural abnormalities before surgery. A Korean study demonstrated that MRI had a positive predictive value of 86% for tumor localization, with intraoperative findings confirming tumors in 77.8% of cases [59]. However, approximately 18% of CD cases have microtumors that are too small to be detected by standard MRI [60]. To improve detection, dynamic contrast-enhanced MRI is recommended over conventional T1- and T2-weighted imaging [61]. That method increases the likelihood of identifying tumors in the far anterior or posterior of the pituitary gland. Further refinements in imaging techniques, such as using half-dose contrast MRI and spoiled gradient recalled echo with 1-mm sections, have also been proposed to enhance the detection of microtumors [62]. Low doses of gadolinium contrast improve differentiation between normal pituitary tissue and tumors, making MRI a crucial tool for surgical planning. Recent studies, including a Korean study using ¹⁸F-fluorodeoxyglucose (FDG) positron emission tomography/computed tomography (PET/CT) have explored alternative imaging modalities for pituitary tumor localization [63]. However, a recent study found that dexamethasone suppression did not significantly improve the localization of ACTH-secreting pituitary tumors using ¹⁸F-FDG PET/CT. Therefore, PET/CT is primarily used to evaluate patients with suspected EAS when biochemical evidence of ACTH-dependent hypercortisolism exists but no pituitary tumor is identified on MRI. In that context, PET/CT helps localize the ectopic source of ACTH production.

1. Clinical context

IPSS is a critical diagnostic tool for differentiating CD from EAS [64]. In Korean studies, IPSS has shown high diagnostic accuracy, with a 100% positive detection rate for CD in a cohort of Korean patients [59]. This supports the continued reliance on IPSS as a gold standard test, particularly when MRI fails to localize a tumor. The procedure involves comparing ACTH levels in blood collected from the bilateral IPSS and a peripheral vein before and after stimulation with either desmopressin or CRH. Desmopressin (administered 10-μg intravenously [IV] over 15 seconds) is an alternative to CRH (100-μg IV), with studies showing comparable efficacy [65].

A baseline IPS/peripheral ACTH ratio >2 is indicative of CD. A post-stimulation IPS/peripheral ACTH ratio >3 strongly suggests CD. Lateralization (a post-stimulation ACTH ratio >1.4 between the two sinuses) suggests the tumor’s location, though this finding correlates with surgical biopsy results in only 38.9% of Korean cases [59]. False negatives can occur due to anatomical variations in venous drainage or procedural technicalities [66]. To ensure catheterization accuracy, prolactin measurement is recommended [67]; an IPS-to-peripheral prolactin ratio >1.8 confirms successful sampling. Additionally, a normalized prolactin-adjusted ACTH ratio >1.3 supports a CD diagnosis, whereas a ratio <0.8 suggests EAS. The KES guidelines recommend checking IPS-to-peripheral prolactin ratios to confirm catheterization. Reports from Japan have demonstrated the usefulness of IPSS in patients with MRI-negative tumors, and prolactin adjustment has been shown to aid in distinguishing between CD and EAS [68]. Although IPSS plays a key role in differentiating between CD and EAS, both countries acknowledge that its accuracy in determining tumor lateralization is limited.