Ach Taieb, Dhaffar Rim, Ammar Asma, Ghachem Aycha, Halloul Imen, Saafi Wiem, El Fekih Hamza, Saad Ghada, Hasni Yosra, Zaouali Monia
Department of Endocrinology, University Hospital of Farhat Hached, Sousse 4031, Tunisia.
Faculty of Medicine of Sousse, University of Sousse, Sousse 4000, Tunisia.
Medicina (Kaunas). 2025 Jul 19;61(7):1303. doi: 10.3390/medicina61071303.
Adrenal insufficiency (AI) is an endocrine disorder characterized by inadequate cortisol production, leading to non-specific symptoms that delay diagnosis. The Low Dose Synacthen Test (LDST) is commonly used to evaluate adrenal function, but traditional cortisol cut-offs may not accurately reflect adrenal function in all patients. This study aims to identify baseline cortisol cut-offs to accurately rule in and out AI, reassess the value of cortisol increment during LDST, and evaluate the accuracy of 30 and 60 min cortisol measurements in diagnosing AI. We conducted a cross-sectional analysis of patients who underwent LDST at Farhat Hached University Hospital. Diagnostic accuracy of baseline cortisol levels and cortisol increment was assessed using ROC curve analysis to determine optimal cut-offs for predicting LDST outcomes. Among 163 patients (mean age 42.9 years, 63% female), baseline cortisol ≤ 5.35 μg/dL had 100% specificity but 41.5% sensitivity for LDST failure. Conversely, baseline cortisol ≥ 12.4 μg/dL had 100% sensitivity with 45.9% specificity. Single measurements at 30 and 60 min correctly classified 92.64% and 93.87% of cases, respectively. ROC analysis of 30 and 60 min cortisol increments showed high diagnostic accuracy (AUC 0.923 and 0.914, respectively). The optimal cortisol increment cut-off was 6.35 μg/dL for ruling in AI (99% specificity). We propose a novel AI diagnostic algorithm based on a single 30 min cortisol measurement, complemented by revised baseline cortisol cut-offs and cortisol increment as additional criteria. This approach may enhance diagnostic accuracy and minimize unnecessary testing, warranting further clinical validation.
肾上腺功能不全（AI）是一种内分泌疾病，其特征为皮质醇分泌不足，会导致出现非特异性症状，从而延误诊断。低剂量促肾上腺皮质激素试验（LDST）常用于评估肾上腺功能，但传统的皮质醇临界值可能无法准确反映所有患者的肾上腺功能。本研究旨在确定基线皮质醇临界值，以准确判断AI的存在与否，重新评估LDST期间皮质醇增加值的价值，并评估30分钟和60分钟时皮质醇测量值在诊断AI中的准确性。我们对在法哈特·哈谢德大学医院接受LDST的患者进行了横断面分析。使用ROC曲线分析评估基线皮质醇水平和皮质醇增加值的诊断准确性，以确定预测LDST结果的最佳临界值。在163例患者（平均年龄42.9岁，63%为女性）中，基线皮质醇≤5.35μg/dL对LDST结果为阴性的特异性为100%，但敏感性为41.5%。相反，基线皮质醇≥12.4μg/dL的敏感性为100%，特异性为45.9%。30分钟和60分钟时的单次测量分别正确分类了92.64%和93.87%的病例。对30分钟和60分钟时皮质醇增加值的ROC分析显示诊断准确性较高（AUC分别为0.923和0.914）。用于判断AI存在的最佳皮质醇增加值临界值为6.35μg/dL（特异性为99%）。我们提出了一种基于单次30分钟皮质醇测量的新型AI诊断算法，并辅以修订后的基线皮质醇临界值和皮质醇增加值作为附加标准。这种方法可能会提高诊断准确性并减少不必要的检测，有待进一步的临床验证。
