Neuroblastoma is the most common solid tumor in childhood and is the most frequent neural crest tumor (NCT). More than 90% of the patients excrete high levels of vanilmandelic acid (VMA) and homovanillic acid (HVA) in the urine. Original biochemical methods for measuring these two metabolites of catecholamines employed a collection of urine for 24 hours to avoid errors related to circadian cycle variations. More recently, attempts have been made to replace the 24-hour collections by random samples (RSs). This has practical advantages particularly for young children. The objective of this study is to assess whether urinary VMA related to urinary creatinine levels can be determined reliably by the method of Pisano et al. from RSs in patients with NCT. The determination of the consumption of VMA in urine stored for prolonged periods of time was also studied. We found a good correlation between the values of metabolites of catecholamines in RSs compared with 24-hour urine collections. There was consumption of VMA in urine samples after storage. We conclude that determination of VMA in RSs of urine by Pisano's method may identify NCT production of catecholamines and that the consumption of these catecholamines is an important factor to consider in the interpretation of values obtained with stored urine specimens.