Cushing's Disease: Data Hints at the Diagnosis

Cushing’s disease (CD), caused by excessive adrenocorticotropic hormone production from a pituitary adenoma, is rare and can be difficult for non-endocrinologists to diagnose. Studies show that CD patients average 2 years of symptoms before receiving the correct diagnosis. Since uncontrolled CD can cause hypertension, glucose intolerance, cardiovascular risks, osteopenia, and other complications, earlier recognition of this disease is crucial.

A study published in January 2016 in the journal Endocrine Practice used a data-mining approach to identify unusual combinations of common clinical conditions that might point to Cushing’s disease .

Researchers from Partnership for Health Analytic Research, LLC, and Novartis Pharmaceuticals Corporation conducted a retrospective matched case-control study using a nation-wide health insurance claims database. In consultation with endocrinologists, the researchers identified 47 signs, symptoms, and comorbidities associated with Cushing’s disease . Using rate and relative risk data, researchers selected 10 conditions and 24 condition pairs to evaluate in the validation group.

The study included 3,750 patients with claims for potentially Cushing’s disease-related ICD-9-CM codes within a 5-year period, and 7,500 matched control patients without Cushing’s disease-related claims. Researchers divided subjects equally into development and validation groups.

CD patients were more than 10 times more likely to have 5 conditions than validation group patients: localized adiposity, hirsutism, facial plethora, polycystic ovary syndrome, and abnormal weight gain.

Four pairs of conditions—hypertension/hirsutism, serious infection/adrenal mass, type 2 diabetes mellitus/hirsutism, and uncontrolled type 2 diabetes mellitus/premature menopause—appeared only in the Cushing’s disease group.

Combinations associated with a relative risk of CD greater than 100 included fatigue/hirsutism, hyperlipidemia/adrenal mass, and type 2 diabetes mellitus/adrenal mass.

Patients who had the uncommon pairing of osteoporosis and obesity were 52 times more likely to have Cushing’s disease than others.

Study authors see their work as a starting point for future research, noting that ICD-9-CM codes lack clinical detail. Electronic medical records could provide more information, leading to identification of other high-risk symptom combinations. However, it may be years before this newer data source includes enough patients for analysis, given CD's rarity.

Clinical paradigm developers could incorporate results from this and future studies into clinical tools, improving identification of high-risk patients, and decreasing time to diagnosis.