Impaired brain function improved by l-carnitine in patients with cirrhosis: evaluation using near-infrared spectroscopy

To evaluate the effects of l-carnitine on impaired brain function in patients with liver cirrhosis. We conducted a retrospective cohort study that included sequential 80 liver cirrhosis patients with impaired brain function evaluated using near-infrared spectroscopy (NIRS). Among them, l-carnitine was administered to 48 patients. The NIRS data and blood ammonia level at baseline and after 8 weeks of treatment were compared between patients administered with l-carnitine (l-carnitine group) and those who were not (control group). The NIRS data at baseline were similar between the l-carnitine and control groups (0.04 ± 0.04 vs. 0.04 ± 0.05 mMmm, p = n.s), whereas those in the l-carnitine group (n = 48) were significantly better than that of the control group at 8 weeks of treatment (n = 32) (0.103 ± 0.081 vs. 0.040 ± 0.048 mMmm, p < 0.001). In the l-carnitine group, 35.4% (17/48) of patients had hyperammonemia. The NIRS data of the l-carnitine group at 8 weeks of treatment were significantly improved than that of the control group, irrespective of baseline ammonia levels (0.11 ± 0.09 vs. 0.04 ± 0.05 mMmm, p = 0.005, and 0.10 ± 0.06 vs. 0.02 ± 0.03 mMmm, p = 0.003, for normal baseline ammonia and elevated ammonia levels, respectively). In the multivariate analysis, l-carnitine administration (odds ratio [OR] 3.51, 95% confidence interval [CI] 1.23–9.99, p = 0.019) and baseline NIRS data of ≤ 0.07 mMmm (OR 5.21, 95% CI 1.69–16.0, p = 0.0041) were found as independent significant factors. l-carnitine improves impaired brain function in patients with liver cirrhosis.

Hepatic encephalopathy is defined as a brain dysfunction caused by liver insufficiency or portosystemic shunting 1 and is characterized by a broad spectrum of neurological or psychiatric abnormalities, ranging from subclinical alterations to coma 1 . Overt hepatic encephalopathy was recognized by the detection of marker symptoms, such as disorientation and flapping tremor, because of their excellent interobserver reliability 2,3 . The positron emission tomography (PET) was used to examine brain oxygen consumption and cerebral blood flow and revealed that patients with overt hepatic encephalopathy and those with liver cirrhosis without overt hepatic encephalopathy had a lower cerebral metabolic rate of oxygen and cerebral blood flow than healthy subjects 4 . Impaired cerebral blood flow and cerebral oxyhemoglobin metabolism in hepatic encephalopathy could be improved 5 . Since the cerebral ammonia uptake remains the same between the state of hepatic encephalopathy and recovery 5 , not only the blood ammonia level but also the brain function should be evaluated in patients with hepatic encephalopathy.
Covert hepatic encephalopathy consists of minimal and grade 1 hepatic encephalopathy 1 , could impair the cognitive function, lower work efficiency, reduce the quality of life 6,7 , impair driving skills [8][9][10][11] , or results in poor prognosis [12][13][14] . Some cases of covert hepatic encephalopathy were difficult to evaluate and diagnose, because of Scientific RepoRtS | (2020) 10:13566 | https://doi.org/10.1038/s41598-020-70585-y www.nature.com/scientificreports/ lacking reliable diagnostic clinical signs. Although several screening tools and neuropsychological tests were developed 2,12,15-31 , brain functional imaging methods for patients with covert hepatic encephalopathy were not established in daily clinical settings. Therefore, a readily available and reliable methodology of functional brain imaging is urgently required in not only patients with overt but also with covert hepatic encephalopathy.
Recently, near-infrared spectroscopy (NIRS) was developed as a readily available brain functional imaging technology in daily clinical settings [32][33][34][35][36][37][38] . Due to the advancements of neurological studies, brain function can be evaluated by monitoring dynamic changes of cerebral blood flow according to neurovascular coupling theory, i.e., neurons, astrocyte, and arterial vessel communication [39][40][41] . NIRS could noninvasively measure the regional cerebral oxygenated hemoglobin (oxyhemoglobin), deoxyhemoglobin, and total hemoglobin concentration. The time resolution of NIRS is higher than that of PET and functional magnetic resonance imaging (MRI). NIRS is also portable, does not have any restrictions in posture, and is flexible in setting tasks. Therefore, brain function can be evaluated based on dynamic changes in regional cerebral oxyhemoglobin concentration in response to a given task. This is essential to assess a latent brain function abnormality. NIRS could also evaluate the regional cerebral oxyhemoglobin metabolism and cerebral blood flow by combining with MRI 34 .
Some reports showed the usefulness of NIRS in patients with liver disease 37,38 . Especially, patients with minimal hepatic encephalopathy showed a lower increase of cerebral oxyhemoglobin concentration in response to word fluency task than that of patients with liver cirrhosis without minimal hepatic encephalopathy 37 . In patients with hepatic encephalopathy, the impaired reaction of cerebral oxyhemoglobin concentration might occur due to astrocyte swelling and dysfunction [42][43][44] because some reports showed that astrocyte controls the regional cerebral blood flow and provides energy sources such as ketone body and carnitine to the neuron 39,41 . l-Carnitine is also reported to improve hyperammonemia and hepatic encephalopathy 45,46 , however, whether l-carnitine can improve those impaired neurons, astrocyte, and vascular interaction remains unknown.
Liver cirrhosis patients were generally considered as having secondary l-carnitine insufficiency. Some patients received l-carnitine replacement therapy by the attending physician's discretion. In all the patients with cirrhosis, l-carnitine therapy is covered by a national insurance system in Japan. Although the definition of carnitine insufficiency in liver cirrhosis patients remains unclear, the symptoms of carnitine deficiency are thought to be muscle cramps, general fatigue, hepatic encephalopathy, and hyperammonemia in liver cirrhosis patients. Therefore the liver cirrhosis patients who suffered from these symptoms were considered to have a possibility of carnitine deficiency. Some attending physicians advised those patients who had impaired brain function to receive l-carnitine supplementation, but some patients did not want to receive it mainly because of avoiding polypharmacy and the cost of medicine.
Therefore, this study aimed to elucidate the effects of l-carnitine on impaired brain function in patients with liver cirrhosis using NIRS.
In the l-carnitine group, 74.2% (23/31) of patients with impaired brain function and with normal ammonia levels had improved NIRS data at 8 weeks of treatment (Fig. 3).
No patient had adverse effects of l-carnitine during the study period.

Discussion
Brain function is recently thought to be an interaction among neurons, astrocytes, and cerebral arterial vessels 39,40 . Astrocyte especially controls the local cerebral blood flow to supply neurons with oxygen and essential nutrients depending on the neural activity 39,41 . The pathophysiology of hepatic encephalopathy is associated with impaired interaction between astrocytes, neurons, cerebral blood flow, and cerebral oxygen consumption 4,5,43,[47][48][49] . This neurovascular coupling could be elucidated by monitoring the dynamic changes of the regional cerebral arterial blood flow. Based on this perspective, brain functional imaging modalities such as PET, SPECT, functional MRI, and NIRS are thought to be useful 50 . NIRS is a portable, easy to set, and has a high time resolution of 0.1 s and is recently reported to be a useful tool to evaluate patients with impaired brain function with liver cirrhosis 37 .
In the present study, NIRS was used to evaluate the effects of l-carnitine on brain activity, which improved after 8 weeks of l-carnitine treatment in patients with cirrhosis. l-Carnitine transports long-chain free fatty acids into the mitochondria, subsequently producing ATP by β-oxidation and TCA cycle and metabolizing ammonia by activating the urea cycle via N-acetylglutamate. Acyl CoA transforms into acetyl CoA through β-oxidation. l-carnitine synthesizes acetylcarnitine from acetyl CoA. Further, acetylcarnitine facilitates acetylcholine production in the striatum and hippocampus 51 .
Therrien et al. reported that l-carnitine reduces the concentration of ammonia and lactate in the cerebrospinal fluid in a study of portocaval-shunted rats 52 . The brain avoids ammonia toxicity by converting α-ketoglutarate into glutamate and glutamine synthesis. In another clinical study, l-carnitine treatment for 6 months reduced blood ammonia levels and improved NCT-B scores in patients with liver cirrhosis with covert hepatic encephalopathy 53 . Another study showed that 45.8% (11/24) of patients with liver cirrhosis with minimal hepatic encephalopathy were ameliorated by l-carnitine administration for 3 months 54 .
In a randomized, double-blind placebo-controlled trial, the 90-day l-carnitine treatment (2,000 mg twice a day) improved hyperammonemia and hepatic encephalopathy 45 . A retrospective analysis of 34 patients with overt hepatic encephalopathy found that oral administration of l-carnitine reduced overt hepatic encephalopathy recurrence 55 . These data suggest that l-carnitine may improve overt or covert hepatic encephalopathy.     In the present study, l-carnitine improved the NIRS data and reduced blood ammonia concentration after 8 weeks of treatment. In the l-carnitine group, 58% of patients with impaired brain function with hyperammonemia had ameliorated NIRS data. Moreover, 23.5% of patients in the l-carnitine group had improved NIRS data without decreased blood ammonia concentration. Conversely, in patients with impaired brain function with normal ammonia levels, NIRS data were significantly improved in the l-carnitine group than that in the control group. Therefore, l-carnitine might improve the brain function not only by improving hyperammonemia but also by other mechanisms.
Recently Wang et al. 56 reported that intracellular reactive oxygen species and all amino acids, including glutamine, were increased in the NH 4 Cl-treated human astrocyte; however, these reactions were reduced with l-carnitine treatment. Moreover, in the NH 4 Cl-treated human astrocytes, increased concentration of 3-methyl-2-oxovaleric acid, which was ammonia induced and played a crucial role in neurological impairment, was significantly decreased with l-carnitine co-treatment. Accordingly, l-carnitine might improve hepatic encephalopathy by not only lowering ammonia concentration but also by directly affecting the brain.
There are some limitations in this study. There is a risk of bias due to non-randomized controlled design and retrospective analysis. But we started observation of patients from the date of NIRS. Therefore we retrospectively analyzed the prospective observational data. And patients were sequentially enrolled. The randomized controlled trial is needed.
Although there was not a statistically significant difference, there were more patients who had BCAA supplementation in the l-carnitine group (50%) compared to the control group (28%). However, there was no patient who started BCAA supplementation during the study period. BCAA has been shown to positively affect sarcopenia, which is an independent risk factor for hepatic encephalopathy itself 57 . In the present study, baseline SMI between the two groups was not different, and baseline skeletal muscle mass depletion was not associated with NIRS data improvement at 8 weeks in this cohort. Further studies are needed in this issue.
In conclusion, l-carnitine administration improves impaired brain function in liver cirrhosis patients.

Methods
The brain function of 140 patients with liver cirrhosis was evaluated using NIRS between June 2013 and October 2019. The diagnostic value of impaired brain function was set as an increased reaction of oxyhemoglobin concentration increased to word fluency task at 10 s, i.e., ≤ 0.1 mMmm using NIRS according to a recent report that compared NIRS and EEG abnormality in liver cirrhosis patients 37 . Among them, 80 patients showed impaired brain function. No patients had a history of mental disorders including alcohol abuse and intake of antidepressants or other psychotropic drugs. All patients underwent brain computed tomography or MRI and had no apparent brain structural diseases, such as brain infarction, brain tumor, etc. All patients performed the number connection test A and B using neuropsychological tests 21 . We analyzed baseline skeletal muscle mass index (SMI) with the CT method using the Slice-O-Matic version 5.0 Software program (Tomovision, Montreal, Canada). Skeletal muscle area was measured on the axial image at the level of   58 . The definition of skeletal muscle mass depletion was based on the guideline described by the Japan Society of Hepatology (42 cm 2 /m 2 in men and 38 cm 2 /m 2 in women). A total of 48 patients were administered 250 mg or 500 mg of l-carnitine three times a day for 8 weeks (l-carnitine group), according to the attending physician's discretion, and the remaining 32 patients were not (control group) (Fig. 4). The study was performed in accordance with the Helsinki Declaration and approved by the ethics committee of Musashino Red Cross Hospital. Informed consent was obtained from all participants.
Assessment of brain activity. The brain function was evaluated using NIRS. The regional cerebral oxyhemoglobin concentration was measured using a 52-channel NIRS machine (Hitachi ETG4000, Hitachi Medical Cooperation, Tokyo, Japan, Fig. 5). NIRS captures changes in regional cerebral oxyhemoglobin concentration every 0.1 s. For each of the 52 channels, optic fiber devices are connected to an application probe placed on the patient's scalp. The 52 channels cover the frontal, upper temporal, and anterior parietal lobes of the brain. The reflected light is detected by a probe positioned 30 mm away from the application probe. Changes in oxyhemoglobin concentration could be calculated by measuring the reflected light 32 . In this study, data were measured using 7 channels, which were previously reported to diagnose mental disorders and minimal hepatic encephalopathy (channels 36-38 and 46-49), were selected for the analysis 35,37,59,60 . The mean increase in oxyhemoglobin concentration of these 7 channels at 10 s during word fluency task was measured. Moreover, changes from baseline to 8 weeks were compared between the two groups.
Statistical analysis. Statistical analyses of the clinical data were performed using Fisher's exact probability test, Student's t test, and a paired t test. We conducted univariate and multivariate logistic regression analysis to obtain the factors associated with improved NIRS data. A p-value of < 0.05 was considered statistically significant. All statistical analyses were performed using EZR software version 2.30 (Saitama Medical Center, Jichi Medical University, Saitama, Japan).