Mortality prediction by serum melatonin levels of patients with spontaneous intracerebral hemorrhage

Objective In one study, higher serum melatonin levels have been reported at diagnosis of spontaneous intracerebral hemorrhage (ICH) in non-surviving than in surviving patients. Now, we carried out this study with the aims to explore whether blood melatonin concentrations in the first 7 days of ICH are different in survivor and non-survivor patients and whether are useful in the prediction of mortality. Methods Six Spanish hospitals participated in this observational study of patients with severe supratentorial ICH (defining severe as Glasgow Coma Scale < 9). We determined serum melatonin levels during the first, fourth, and eighth day of severe ICH. Results Surviving (n = 64) compared to non-surviving (n = 53) patients showed lower serum melatonin levels during the first (p < 0.001), fourth (p < 0.001), and eighth day (p < 0.001) of severe ICH. We found in multiple logistic regression analysis an association between serum melatonin levels and 30-day mortality (odds ratio = 8.932; 95% CI = 2.442-32.665; p = 0.001) controlling for midline shift, ICH score, early evacuation of ICH, and glycemia. We found an AUC (95% CI) for the mortality prediction of 0.90 (0.83-0.95; p < 0.001), 0.94 (0.87-0.98; p < 0.001), and 0.90 (0.81-0.96; p < 0.001) by serum melatonin concentrations during the first, fourth, and eighth day. Conclusions In our current study, it appears that novel findings of serum melatonin levels recollected at any moment during the first 7 days of a severe ICH were higher in non-survivor than in survivor patients and could help in mortality prediction.