Purpose: Hypercarotenemia and transaminitis are reported as laboratory features of anorexia nervosa. However, the specificity and sensitivity of an elevation in serum carotene and transaminase are not known. Therefore, a prospective study was undertaken to determine the clinical utility of these serum markers.
Methods: Serum carotene was measured in 46 female adolescents between 13 years and 18 years of age (21 anorexia nervosa, 17 bulimia nervosa, 8 unclassified eating disorders). Findings were compared to levels of carotene in serum samples obtained from similarly aged females with either chronic inflammatory bowel disease (22 Crohn disease, 11 ulcerative colitis) or acute medical symptoms not associated with undernutrition or intestinal inflammation (N = 26), and 21 children of either sex with dyspeptic symptoms.
Results: Serum carotene was elevated in 6/46 (13.0%) females with eating disorders compared with only 2/80 (2.5%) children in the three comparison groups (p < 0.01). Hypercarotenemia was present in 4/21 girls with anorexia nervosa compared with 0/17 females with bulimia nervosa (p = 0.11). Transaminitis was present in 38.5% (AST) and 7.7% (ALT) of eating disorder patients. Liver enzyme abnormalities, however, did not correlate with hypercarotenemia. Transaminitis was also not specific for eating disorders since transaminitis was observed with comparable frequency in the three comparison groups.
Conclusions: These findings confirm that hypercarotenemia is a laboratory feature in some subjects with eating disorders, in particular, anorexia nervosa. The low sensitivity (13.0%) does not provide justification for its use as a screening test. However, in complicated diagnostic settings a serum carotene determination could prove useful because the specificity (97.5%), positive predictive value (75.0%), and negative predictive value (66.1%) of an elevated carotene were high. These data also show that elevated carotene levels are not associated with hepatic abnormalities. Although transaminitis is reported as a laboratory feature of eating disorders, the prevalence of such abnormalities in this study was not higher than in age-matched comparison groups.