Development of a Novel Clinical Screening System for Cushing’s Disease in the United States

Background: Cushing’s disease (CD) lacks a single pathopneumonic symptom and carries clinical overlap with other metabolic syndromes, contributing to diagnostic delay. Lag time increases morbidity and mortality given long-term sequelae of prolonged glucocorticoid exposure. Clinical screening systems (CSS) may be an effective first-line screening tool to identify and refer patients with high pre-test probability of CD for definitive biochemical workup. The Italian and Spanish CSS have been previously defined and validated in Europe. These systems, however, carry poor sensitivity in the United States patient population. Here, we introduce a novel CSS for patients with CD in the United States.

Methods: Data were obtained from the Registry of Adenomas of the Pituitary and Related Disorders (RAPID) skull base consortium database and CDC 2019 National Health Interview Survey (NHIS) datasets. Analyses were performed on 615 patients who underwent endoscopic endonasal surgery for CD between 2003 and 2023 and 31,997 patients representing the general U.S. population. We identified patient factors that reliably distinguished CD patients from the general population to define a novel clinical scoring system referred to here as the RAPID scoring system. Hypertension is the most common comorbidity in CD and is associated with increasing age; we, therefore, focused analyses on defining how this association differed from the general U.S. population. Differential prevalence of hypertension was calculated for males versus females, obese versus non-obese, and ≥ or ≤ 50 years of age. Differential prevalence was defined as the difference in prevalence of the comorbidity between patients with and without CD. Chi-squared was used to assess difference in binary risk stratification between the three scales (Italian, Spanish, and RAPID systems). To assess bias, t-tests were calculated for the mean scores based on sex, race, ethnicity, age, and obesity for each scale. Sensitivities were compared using Cochran’s Q test.

Results: CD was more likely to present with the combination of hypertension and any of the following: age < 50, non-obesity, and male sex. Differential prevalence of hypertension between patients with and without CD was 41.4% for males compared to 29.5% for females, 45.8% for patients < 50, compared to 27.6% for patients ≥ 50, and 28.1% for non-obese patients compared to 23.7% for obese patients. The RAPID scoring system was constructed as follows: +3 hypertension, +1 hypertension plus age < 50, +1 hypertension plus non-obesity, +1 hypertension plus male sex, +2 diabetes mellitus, and +2 for any Cushing’s symptom observed in at least one-fourth of CD patients in the RAPID dataset ([Fig. 1]). Fatigue was excluded due to nonspecificity. Score distribution revealed a notable step-off at a score of 6, which was chosen as the empiric threshold ([Fig. 2]). The RAPID scale demonstrates increased overall sensitivity (81.0%) compared to the Spanish (41.5%) and Italian scales (67.3%; [Figs. 2] [3] [4]; p < 0.001).

Conclusion: We introduce a novel clinical screening system constructed from analyses on large datasets of patients with confirmed CD and patients who are representative of the general U.S. population. The RAPID System expectedly provides greater sensitivity compared to conventional European scoring systems given inherent differences in the patient population.

 

MissingLinkText

No conflict of interest has been declared by the author(s).

Publication History

Article published online:
07 February 2025