The database is well corresponded to the whole population; therefore, loss of follow-up or selection bias were not concerns. Other findings in our study are consistent with previous publications. First, age is still the single strongest precipitating factor for dementia. Other risk factors such as female, diabetes, stroke and SES were illustrated before. Of note, hypertension, hyperlipidemia, history of alcohol intoxication and coronary artery disease were not associated with increased risk of dementia in our study. The possible reasons are that there are still conflicting results in publications, and their effects on dementia may be partly explained by the coexisting comorbidities or CCI. In conclusion, TBI is an independent significant risk factor of developing dementia even in the mild type. The result indicates that more emphasis on the head injury prevention would be worthy. Fat mass and obesity associated protein is an AlkB-like 2oxoglutarate �C dependent nucleic acid demethylase, which exhibits substrate specificity for 3-methylthymidine and 3-methyluracil in single-stranded DNA and RNA. However, recent structural data show that specific hydrogen bond interactions account for the preference of Fto for thymine or uracil over adenine, cytosine or guanine, suggesting that methylated single-stranded RNA, rather than DNA, may be the primary Fto substrate. Sequence analysis also predicted that human Fto and its vertebrate homologs are globular proteins that carry a nuclear localization signal and are unlikely to be targeted to membranes or organelles. The studies of both rodent and human Fto mRNA expression have shown that this protein is expressed in many tissues,, with concentrations especially high in the hypothalamic sites that govern feeding behavior, such as arcuate, paraventricular, dorsomedial and ventromedial nuclei. Although the connection between single nucleotide polymorphisms in the Fto gene and body mass index was established long ago, downstream effects of changes in Fto expression remain unexplored. Animal experiments indicate a relationship between Fto levels and energy metabolism and food intake, doppler echocardiography evaluating aortic velocity impacting body weight. For instance, it has been reported that Fto expression was significantly increased in the hypothalamus of food-deprived and food-restricted rats. Considering that selective modulation of Fto levels in the hypothalamus influences food intake, our goal was to determine the effect of fasting on Fto expression in lateral hypothalamic area, PVN, VMN and ARC, all of which are the regulatory centres of energy homeostasis. It is known that a change in nutritional status primarily affects Fto levels in hypothalamic regions involved in the regulation of energy homeostasis. Studies conducted on mice suggested that fasting induced a decrease in Fto mRNA in ARC. However, experiments on rats showed the opposite. In that respect, our findings are in line with those of Olszewski et al., 
where the enhanced expression of both Fto mRNA and the protein was detected in hypothalamus of 48 h starved rats. Our immunofluorescence results have confirmed the increased Fto expression in some LHA, PVN and MVN neurons. Surprisingly, although it had been unequivocally accepted that Fto is exclusively a nuclear protein, our studies revealed that the majority of Fto was localised in cytoplasm surrounding neuronal nuclei. Bearing in mind that we detected the upregulation of both the Fto mRNA and the protein expression as early as 6 hours after the food had been removed, it is unlikely that the Fto detected in cytosol was just newly synthesized protein not yet translocated into the cell nucleus. If this were the case, the increased amount of Fto would also have been detected within the nuclei of LHA, PVN and MVN neurons after 48 h of food deprivation.