Piecing Together the Adrenal Puzzle

Four Studies that Can Improve Patient Outcomes Since April is Adrenal Disease Awareness Month, Endocrine News is highlighting recent studies that shed light on important advances and insights in the field. These studies, all published in Endocrine Society journals, show how improved diagnostic accuracy, postoperative management, and a better understanding of rare adrenal pathologies offer clinicians tools that could improve patient outcomes. At a Glance • The CAI score improves diagnostic accuracy in ambiguous cases of suspected central adrenal insufficiency (CAI) by integrating morning serum cortisol with key clinical parameters and can help guide treatment decisions. • Two-pronged testing consisting of basal cortisol and cosyntropin levels initiated as early as four weeks post-adrenalectomy can identify patients with adrenal insufficiency, guide glucocorticoid treatment, and obviate unnecessary glucocorticoids. • Although typically benign and nonsecretory, adrenal ganglioneuromas (AGNs) can rarely exhibit secretory properties that mimic other adrenal pathologies, making definitive histopathologic assessment essential; in cases of ACTH-dependent hypercortisolism with an adrenal lesion, secretory AGN should remain on the differential. • MEN1-associated adrenocortical carcinoma requires comprehensive hormonal evaluation, regular follow-up for adrenal lesions, and routine screening of asymptomatic mutation carriers for early detection and improved outcomes. Given the adrenal glands’ enormous hormonal reach — governing inflammation, blood pressure, stress response, and more — when they are dysfunctional, the consequences are serious. According to the Endocrine Society’s Endocrine Facts and Figures , the prevalence of primary adrenal insufficiency is estimated at 40 to 100 cases per million in the United States, while Cushing syndrome affects an estimated 8 people per million in those under age 65 years. Rarer still, Cushing disease occurs in roughly 2.3 to 2.7 cases per million per year, and multiple endocrine neoplasia type 1 (MEN1) affects an estimated 3 to 10 people per 100,000. Those numbers may look modest, but behind each one is a clinical journey often fraught with challenges. Two research studies advance our understanding of cortisol testing and postoperative management, while two case reports serve as vivid reminders that the adrenal gland can still confound even the most experienced clinicians. Central Adrenal Insufficiency In “ CAI Score for the Diagnosis of Central Adrenal Insufficiency ,” published in the Journal of the Endocrine Society in February, Mussa H. Almalki, MBBS, MHSc, of the College of Medicine of Alfaisal University in Riyadh, Saudi Arabia, and team have truly moved the needle forward when it comes to day-to-day, in-clinic management. As to what prompted this study, Almalki credits the well-known frustration that comes from wanting to help patients but being hemmed in by existing clinical parameters. “We often see patients who we suspect might have central adrenal insufficiency (CAI) — perhaps they have a pituitary tumor, have had head trauma, or have other hormone deficiencies,” he recounts. “We order a morning cortisol test, hoping for a clear answer. But so often, the result comes back in what we call the ‘gray zone’—typically between 4 and 18 µg/dL. It’s not low enough to confidently diagnose CAI, but it’s not high enough to rule it out.” This dilemma generally creates the need to order a dynamic test (e.g., short Synacthen test [SST]), but this, says Almalki, is time-consuming, requires patients to come to a dedicated unit, can be unpleasant, and is not a perfect test itself. “We realized we needed a better way to stratify risk in these ‘gray zone’ patients before deciding on next steps. We wanted to see if we could combine the information we already had — like the specific cortisol level, the presence of other pituitary issues, and imaging findings — to build a more sophisticated tool than just looking at the cortisol number in isolation. The goal was to help clinicians make a more informed, data-driven decision about who truly needs that dynamic test.” For their retrospective single-center study from a Riyadh tertiary referral center, the team enrolled 341 adults with suspected CAI and indeterminate morning cortisol levels, using the SST as the reference standard for diagnosis. They developed and validated a predictive scoring system that integrates morning cortisol levels, pituitary hormone deficits, tumor size, sex, and treatment history to help stratify CAI risk in the diagnostically challenging “gray zone” where cortisol results alone are inconclusive. Where traditionally, the “gray zone” is defined as about 3 to 15 µg/dL, the team deliberately expanded it to 4 to 18 µg/dL to be useful in the real world, where different labs use different cutoffs. “But more importantly,” explains Almalki, “we know a cortisol of 5 µg/dL isn’t the same as 15 µg/dL, even if both are ‘gray.’ By widening the range, we could let the data show us how risk changes as cortisol drops, rather than forcing it into arbitrary boxes.” The resulting tool — the CAI score — which also incorporates a machine learning model and is freely available as a web-based application, demonstrated stronger diagnostic accuracy than morning cortisol alone. Along the way, the researchers encountered a couple of surprises. The first was what Almalki calls “the sheer power of pituitary hormone deficits.” In their model, having three or more additional hormone deficiencies was a very strong predictor (odds ratio >35). “This really drove home the point that CAI is very rarely an isolated event. It’s often a sign of more widespread pituitary damage. The health of the pituitary gland as a whole is a massive clue to corticotroph function,” he says. The second was the comparatively modest role tumor size played. “While larger tumors did increase the risk, size wasn’t as powerful a predictor as the number of other hormone deficits. This suggests that it’s not just the size of the tumor, but how it’s impacting the function of the surrounding healthy pituitary tissue — as evidenced by the other hormone losses — that truly matters for CAI risk,” explains Almalki. “Endocrinology is a field defined by complex, interacting feedback loops. A single lab value rarely tells the whole story. AI-assisted tools are perfectly suited to integrate multiple data points — labs, imaging, symptoms, other diagnoses — and recognize patterns that are too subtle or complex for the human brain to consistently process. I see these tools not as replacing the endocrinologist, but as powerful allies that will handle the ‘noise’ and allow us to focus on the ‘signal.’” — Mussa H. Almalki, MBBS, MHSc, College of Medicine, Alfaisal University, Riyadh, Saudi Arabia The future for the CAI score certainly looks promising, but two things need to happen for its widespread adoption, according to Almalki: “First, external validation — seeing the model perform consistently across different hospitals and patient populations. Second, demonstrating practical value. If we can show that using the score reduces unnecessary testing, saves money, and doesn’t compromise patient safety, that creates a compelling case for integration into electronic health records or clinical workflows.” The team is not actively pursuing additional related studies, because, as Almalki puts it, “My immediate focus is shifting toward clinical application and direct patient care, rather than driving the next prospective study. I’m happy to leave the door open for other researchers to pick this up and validate it if they see value in it.” As for the importance of integrating AI-assisted tools in medicine, these researchers emphasize such tools are intended to support rather than replace clinical judgment. “Endocrinology is a field defined by complex, interacting feedback loops. A single lab value rarely tells the whole story. AI-assisted tools are perfectly suited to integrate multiple data points — labs, imaging, symptoms, other diagnoses — and recognize patterns that are too subtle or complex for the human brain to consistently process. I see these tools not as replacing the endocrinologist, but as powerful allies that will handle the ‘noise’ and allow us to focus on the ‘signal,’” Almalki says. Finally, for those clinicians similarly frustrated by how to manage a condition in the face of ambiguous results, Almalki has actionable advice: “Stop looking at that gray-zone cortisol in isolation. It’s just one piece of the puzzle. Ask yourself: How low is it? Do they have other hormone issues? What does their MRI show? The CAI score just helps you put those pieces together quickly. A low score might save a patient an unnecessary test. A high score tells you to stop messing around and act. It’s free, it’s fast, and you can use it right now on the website.” [Go to https://cai-predictor.streamlit.app/ .] Post-Adrenalectomy Adrenal Insufficiency In “ Cortisol Testing to Diagnose Adrenal Insufficiency Following Adrenalectomy for Mild Autonomous Cortisol Secretion ,” published just last month in The Journal of Clinical Endocrinology & Metabolism , a team of researchers led by Oksana Hamidi, DO, MSCS, associate professor of medicine of the University of Texas Southwestern Medical Center in Dallas, Texas, and corresponding author Irina Bancos, MD, MSc, professor of medicine and Adrenal Lab Principal Investigator of the Mayo Clinic in Rochester, Minn., sought contemporary data using standardized cortisol thresholds and modern assays on the true prevalence and duration of adrenal insufficiency after adrenalectomy for mild autonomous cortisol secretion (MACS). “Prior studies were heterogeneous, and there remains uncertainty about the optimal postoperative testing strategy,” says Bancos. “Specifically, we aimed to clarify three issues: (1) how often adrenal insufficiency occurs after unilater…