P1240EVALUATION OF COGNITIVE FUNCTION IN HEMODIALYSIS PATIENTS USING MAGNETIC RESONANCE SPECTROSCOPY: FIVE YEARS FOLLOW UP RESULTS

Abstract Background and Aims In Japan, the elderly population is increasing remarkably, and dialysis patients are aging as usual. According to statistics from the Japanese Society for Dialysis Therapy, at the end of 2017, the incidence of HD patients is estimated to be 3 per 1,000 population, and by the end of 2020, the average age will be over 70 years. Therefore, early diagnosis of cognitive impairment is an important issue. With the aging of dialysis patients, the number of cases showing cognitive dysfunction increase in addition to the decline of physical strength. The problem of the onset of dementia involves many difficulties in medical treatment and nursing. Magnetic resonance spectroscopy (MRS) has been progressing from 1970s to evaluate neurological functions by measuring metabolites in the brain non-invasively. There are few reports using MRS for dialysis patients. In this study, we investigated the brain metabolites of hemodialysis (HD) patients with or without cognitive impairments using MRS and evaluated its usefulness for the diagnosis of cognitive disorder. Method A Toshiba MR device of 1.5 T was used. PRESS sequence was used to acquire water-suppressed 1H-MRS. Timing was TR/TE 2000/25 ms. Three kinds of brain metabolites, namely N-acethylaspartate (NAA), creatine (Cr) and mioinositol (MI) in the posterior cingulate gyrus were measured for 25 healthy adults (Cont group, 44±16 y.o.) and 84 HD patients (HD group, 74±11 y.o.), and ratios of NAA/Cr, MI/Cr and MI/NAA were calculated. The concentration of each metabolite was analyzed using LC model. HD patients were classified into three groups, namely normal cognitive function group (HD-N, n=25, 72±16 y.o.), mild cognitive impairment (HD-M, n=29, 74±9 y.o.) and dementia (HD-D, n=30, 79±8 y.) using MMSE test. Also, sequential changes of the brain metabolites were evaluated for 13 patients with worse cognitive function prospectively. Results HD patients showed a significant decrease of NAA and increases of MI and MI/NAA ratios compared to those of Cont group, suggesting that some metabolic abnormalities were inducted in HD. With a detailed classification of cognitive function in HD patients, NAA/Cr ratios were 1.69±0.17, 1.57±0.15, 1.71±0.20 and 1.54±0.22 in Cont, HD-N, HD-M and HD-D groups, respectively, and was significantly lower even in HD-N group than that of Cont group. MI/Cr ratios were 0.78±0.21, 0.90±0.21, 0.95±0.28 and 1.02±0.27 in Cont, HD-N, HD-M and HD-D groups, respectively, and those of HD-N/-M/-D were significantly higher than that of Cont group. Also, the value of HD-D was significantly higher than those in the other groups. MI/NAA ratios were 0.46±0.13, 0.56±0.17, 0.54±0.16 and 0.66±0.15, in Cont, HD-N, HD-M and HD-D groups, respectively. Again, those of HD-N/-M/-D were significantly higher than that of Cont group. HD-D group was highest among the HD patients. In the prospective study, dementia progressed in 10 of 13 HD patients who were observed more than 5 years. The MI/NAA ratio increased in the patients with dementia progression (from 0.58±0.11 to 1.24±0.17) while that value of the patients without dementia progression showed no changes (from 0.51±0.14 to 0.55±0.18). Conclusion These result suggest that the measurement of metabolic fluctuation in the brain using MRS is useful for the diagnosis of cognitive function in HD patients. The MI/NAA value is a strong candidate for a predictive biomarker of dementia progression. In the future, research and development of measurements of various parts of the brain and their integration to show changes in the whole brain are desired.