Overlooked Biotin Interference in Children

Introduction The earliest reports of “biotin interference” are in neonates with rare biotinidase deficiency, who are prescribed oral biotin. Some infants suspected of having a metabolic disorder may also be started on a “cocktail” of supplements including biotin pending definitive genetic diagnosis. Like all paediatric hospitals we have biotin treated patients and the laboratory performs immunoassays, some with potential for biotin interference. We describe a paediatric patient with congenital kidney disease where falsely high 25-OH vitamin D and falsely low PTH (Roche Cobas) indicated vitamin D toxicity when the patient was vitamin D deficient. Case Details The 15 month old infant has complex medical problems including congenital kidney disease. Serum creatinine was stable at 125 μmol/L (10–30 μmol/L) and albumin 25 g/L (29–45 g/L). At 12 months of age the patient experienced an acute deterioration possibly due to an underlying metabolic disorder and commenced oral biotin. The patient’s 25-OH vitamin D (Roche) levels were >175 and 143 nmol/L so the team diagnosed vitamin D toxicity. Two months earlier the patient’s results were 67–85 nmol/L (>50 nmol/L). The patient’s PTH was between 13–29 pmol/L (Roche; 1.3–6.8 pmol/L). Upon an alert to biotin intake, samples were re-analysed and 25-OH vitamin D was verified as 21 nmol/L (Diasorin Liaison XL) and PTH was higher than originally reported (69 pmol/L; Abbott Architect). Conclusion In chronic renal failure PTH is elevated and 25-OH vitamin D is usually in the deficiency range. Extremely elevated 25-OH vitamin D is uncommon but occurred twice, and was supported by a PTH not as high as expected for chronic kidney disease. Analytical interferences which impact multiple immunoassays is rare but includes biotin therapy, so these children are vulnerable to erroneous laboratory results. No laboratory is completely exempt from biotin-based interference, so paediatric facilities should consider a system-wide approach to safeguard against this predictable error in a known high risk population.