The aim of this study was to evaluate whether nocturnal melatonin concentration (NMC) and urinary 6-sulphatoxymelatonin (aMT6s) excretion can predict melatonin status in patients with severe sepsis in the pediatric intensive care unit (PICU). Blood samples for the determination of NMC were obtained from each patient at 3 a. m. Urine samples for the determination of aMT6s excretion were obtained from each patient at 12 h intervals. We obtained 89 blood and 178 urine samples from 23 septic patients, and 52 blood and 104 urine samples from 13 non-septic patients. The NMC of septic patients in a state of septic shock was significantly higher than that of septic patients not in septic shock (p=0.017) and those of non-septic patients (p=0.019). In contrast, there was no significant difference for nocturnal (NaMT6s) and total aMT6s (TaMT6s) excretion between septic patients with and without septic shock and non-septic patients (p>0.05). The NMC was significantly higher in septic patients in shock with and without hepatic dysfunction (HD) than in non-septic patients (p=0.004 and p=0.024, respectively). NaMT6s and TaMT6s excretion was significantly lower in septic patients with HD than in septic patient without HD (p=0.040 and p=0.029, for NaMT6s and TaMT6s, respectively). Our results showed that an elevated NMC may not reflect an increased MT production in septic patients in septic shock. It seems that, to evaluate the melatonin status of septic PICU patients, it is necessary to collect both serum and urine samples.