Ordering pattern and performance of biochemical tests for diagnosing pheochromocytoma between 2000 and 2008

• Published in: Endocrine practice Vol. 15; no. 4; pp. 313 - 321
• Main Author: Yu, Run
• Format: Journal Article
• Language: English
• Published: United States Elsevier Limited 01.05.2009
• Subjects: Adolescent; Adult; Aged; Aged, 80 and over; Catecholamines - urine; Female; Humans; Male; Metanephrine - blood; Metanephrine - urine; Middle Aged; Pheochromocytoma - diagnosis; Pheochromocytoma - urine; Predictive Value of Tests; Vanilmandelic Acid - urine; Young Adult
• ISSN: 1530-891X; 1934-2403
• DOI: 10.4158/ep09032or

To examine what tests are ordered by physicians for pheochromocytoma diagnosis and how those tests perform in modern clinical practice. In this case series, electronic medical records of patients seen between January 2000 and July 2008 at a large academic hospital in Los Angeles, California, were queried, and patients older than 15 years who underwent any 1 of 5 tests for pheochromocytoma (measurement of plasma catecholamines, plasma fractionated metanephrines, urinary catecholamines, urinary metanephrines, or urinary vanillylmandelic acid) were identified. Because testing was performed in various reference laboratories, test results were classified into 1 of 3 categories: (a) markedly elevated, (b) moderately elevated, or (c) normal. Patient demographics, clinical history, test results, imaging study findings, and pathology records were reviewed. A total of 3980 tests were ordered for 1898 patients. Pretest probability was 2.2% (based on 681 patients in whom pheochromocytoma was confirmed or excluded), and hypertension was the most common indication for testing. The number of patients tested and the number of tests ordered increased over the years. The ordering pattern stabilized since 2006 when urinary metanephrines, urinary catecholamines, and plasma metanephrines were ordered more frequently. Sensitivity was highest for urinary metanephrines and vanillylmandelic acid, specificity was highest for vanillylmandelic acid and urinary catecholamines, and positive likelihood ratio was highest for vanillylmandelic acid. Positive predictive value for markedly elevated test results was 39% to 83%, while that for moderately elevated test results was only 2% to 14%. Ordering pattern and test performance differ significantly from those recommended and reported by large centers. The best testing strategy should incorporate local experience. Categorizing test results as markedly elevated, moderately elevated, and normal is important for result interpretation.