2019 Volume 25 Issue 1 Pages 107-113
D-Glucosamine (GlcN) has been widely consumed as a dietary supplement because of its health benefits. However, limited information exists on the adverse effects of high-dose GlcN treatment. In this study, we investigated the effect of dietary 4 % GlcN hydrochloride on serum and cecum parameters in rats. Growth and food intake were unaffected, but serum levels of ammonia and ethanol significantly increased by GlcN (+21 % and +12 %, respectively). No changes in serum parameters, including AST, ALT, LDH, γ-GTP (indices of liver damage), and urea, were found. The GlcN intake significantly increased the weights of cecal contents (+115 %). Furthermore, supplemental GlcN significantly elevated the levels of ammonia and ethanol (+27 % and +93 %, respectively) and the number of total bacteria (+79 %), when expressed per gram of cecal contents. Our results suggest that high-dose GlcN causes adverse effects by increasing ammonia and ethanol levels and by bacterial overgrowth.
D-Glucosamine (2-amino-2-deoxy-D-glucose, GlcN) supplements have been widely used to relieve arthritic symptoms (Houpt et al., 1999). Several beneficial properties of GlcN have been reported, such as atheroprotective (Largo et al., 2009) and anti-obesity effects (Huang et al., 2015). For example, the consumption of GlcN hydrochloride (HCl) at a dose of 500 mg/kg attenuated the increases of intra-abdominal fat, serum leptin levels, and insulin resistance induced by a high-fat diet in rats (Barrientos et al., 2010). Additionally, the consumption of GlcN sulfate, applied at a dietary level of 0.1 %, was found to ameliorate colitis by suppressing the activation of NF-κB and the related inflammatory responses (Bak et al., 2014). GlcN intake lowered systemic inflammation and altered other pathways in healthy overweight individuals (Navarro et al., 2015). Epidemiologic evidence is present that GlcN is associated with a reduced risk of colorectal cancer (CRC) (Kantor et al., 2016). Additionally, recent studies revealed the anti-aging effect of 1 % GlcN HCl (Gueniche and Castiel-Higounenc, 2017). Therefore, accumulating evidence indicates the health effects of the application of GlcN as a dietary supplement. Meanwhile, feeding of high-dose GlcN HCl (5 % w/v) in the drinking water for 10 weeks exerted an adverse effect by promoting endoplasmic reticulum stress, including hepatic steatosis, and accelerating atherogenesis in apoE−/− mice (Beriault et al., 2011). Long-term consumption of GlcN HCl (50 mg/kg body wt/day) for 20 weeks was reported to increase circulating cholesterol concentrations in mice (Tannock et al., 2006). However, the number of studies on the adverse effect of high doses of GlcN on the colonic luminal environment is still limited. Thus, in this examination, we investigated the blood and cecum parameters associated with diseases in rats fed high-dose GlcN (4 % GlcN HCl-containing diet).
Animals and diets A total of 16 male growing (3-week-old) Wistar rats were purchased from the Hiroshima Laboratory Animal Center and maintained according to the “Guide for the Care and Use of Laboratory Animals” of Hiroshima University (Hiroshima, Japan). This study was approved by the Ethics Committee of Hiroshima University (approval No. C15-12). The rats were housed individually in an air-conditioned room at 23 °C–24 °C under a 12-hour light/dark cycle (lights on from 08:00–20:00). Following acclimatization with a non-purified commercial rodent diet (MF, Oriental Yeast, Tokyo, Japan) for 7 days, the rats were randomly assigned to a control diet group or an experimental diet group (n = 8 rats per group). The composition of the experimental diets was based on a standard diet (Table 1). The experimental diet included 4 % (w/w) GlcN HCl (>99.0 %, Wako Pure Chemical Industries, Ltd., Osaka, Japan). Equal amounts of the diets were incorporated daily into food cups at 18:00 (8, 15, 18, 20, 22, and 24 g on days 1–3, 4–8, 9–4, 15–20, 21–35, and 36–56, respectively) to prevent differences in the food intake of the animals. All food was consumed daily until the next day's food was served. The weight of the spilled diets was recorded daily and appropriately incorporated in the calculation of food intake. Fecal pellets were collected over the last 2 days, stored at −35 °C, and then freeze-dried and milled. The milled samples were stored at −35 °C until subsequent analysis of IgA and mucins. At the end of the feeding period, the rats were sacrificed by decapitation after inhalation exposure to isoflurane (dissection time 13:00–15:00). Blood was collected, and the serum was separated by centrifugation at 2,000 × g for 20 min at 4 °C and stored at −70 °C. Liver, epididymal adipose tissue, and perirenal adipose tissue were removed and weighed. The cecum was immediately excised, and its contents were completely removed, weighed, and stored at −80 °C until further analysis. A portion of the contents was used for DNA extraction immediately after its collection.
| Experimental diets (%, w/w) | ||
|---|---|---|
| Control | GlcN3 | |
| Glucosamine HCl | 0 | 4 |
| L-Cystine | 0.3 | 0.3 |
| Casein | 20 | 20 |
| Vitamin mixture1 | 1 | 1 |
| Salt mixture2 | 3.5 | 3.5 |
| Cellulose | 5 | 5 |
| Sucrose | 20 | 20 |
| Corn oil | 5 | 5 |
| Corn starch | 45.2 | 41.2 |
Measurements
Serum parameters The serum parameters were analyzed using a Beckman Coulter AU480 analyzer (Beckman Coulter, Krefeld, Germany), which is an automated chemistry instrument for turbidimetric, spectrophotometric, and ion-selective electrode measurements. Briefly, 200 µL of serum was used to measure aspartate aminotransferase (AST), alanine aminotransferase (ALT), lactate dehydrogenase (LDH), gamma-glutamyl transpeptidase (γ-GTP), triglyceride, total cholesterol, low-density-lipoprotein (LDL) cholesterol, free fatty acids, albumin, urea, uric acid, ammonia, Ca, Fe, inorganic phosphorus, Mg, glucose, and lactic acid according to the manufacturer's instructions.
Serum and cecum ethanol and cecum ammonia The levels of ethanol in the serum and cecal contents were determined using an enzymatic kit. Briefly, 100 mg of each cecal content was homogenized with 500 µL of deionized water and filtered with the 10 kDa Millipore Ultra free-MC PLHCC HMT Centrifugal Filter Device (Millipore, Bedford, MA, USA). Ethanol levels were quantified using a UV-method for determination of ethanol (Boehringer Mannheim/R-Biopharm, Darmstadt, Germany), following the manufacturer's instructions. Cecum levels of ammonia were measured with the Ammonia-Test Wako Kit (Wako Pure Chemical Industries, Ltd., Japan). The cecal contents were deproteinized with the deproteinizing reagent solution and centrifuged at 2,000 × g for 5 min. The levels of ammonia in the supernatant were determined using the kit.
Cecum microflora Bacterial genomic DNA was isolated from the cecal digesta using the UltraClean™ Fecal DNA extraction kit (MO BIO Laboratories, Carlsbad, CA, USA) according to the manufacturer's instructions. According to the method described elsewhere (Yang et al., 2017), bacterial groups were quantified by real-time quantitative PCR (qPCR) using the StepOne Real-Time PCR System (Applied Biosystems, Foster City, CA, USA). Real-time qPCR was performed in a reaction volume of 20 µL containing 10 µL of SYBR qPCR mix and 2 µL of the enhancing solution (Toyobo Co., Ltd., Osaka, Japan), 1.6 µL of the forward and reverse primer mixture (200 nM each) described elsewhere (Yang et al., 2017), 2 µL of cecal DNA sample, and 4.4 µL of autoclaved distilled water. After initial denaturation at 95 °C for 30 s, 40 PCR cycles were performed as follows: denaturation at 95 °C for 5 s, annealing at 55 °C for 30 s, and extension at 72 °C for 15 s. The fluorescent products were detected at the last step of each cycle. Melting curve analysis was conducted after the amplification to distinguish the targeted PCR product from the non-targeted one. Data were analyzed by the second derivative maximum method of the StepOne Real-Time PCR Software. Plasmid copy number/µL was determined for standard plasmid solution [(cut standard plasmid mixture ng/µL) × (molecules bp/1.0 × 109 ng) × (1/660 DNA length bp/plasmid) = plasmid copies/µL]. Real-time qPCR products were run using serial dilutions of each standard mixture to compare threshold cycle number with copy numbers of the target sequence and to generate standard curves for the quantification of unknown samples. Typically, standard curves were linear across five orders of magnitude (R2 > 0.99).
Fecal immunoglobulin A and mucins Total immunoglobulin (IgA) concentrations in feces were measured using an enzyme-linked immunosorbent assay (ELISA) quantitation kit (Bethyl Laboratories, Montgomery, TX, USA). Briefly, 100 mg of freeze-dried feces was homogenized in 4 mL of ice-cold phosphate buffer saline (pH 7.2) containing 0.1 mg/mL of soybean trypsin inhibitor (Wako Pure Chemical Industries), 50 mM ethylenediaminetetraacetic acid, and 1 mM phenylmethylsulfonyl fluoride. The homogenized solution was incubated at 4 °C overnight and then centrifuged at 9,000 × g for 10 min. The supernatant was collected and stored at −80 °C until subsequent analysis of IgA. Before analysis, the collected supernatant was diluted 200 times, and total fecal IgA was measured by using the ELISA Quantitation Kit according to the manufacturer's protocol.
Fecal mucins were extracted as described by Bovee-Oudenhoven et al. (1997). Briefly, freeze-dried feces were suspended in 20 volumes of phosphate buffered saline, mixed, and immediately incubated in a shaking water bath at 95 °C for 10 min to denature glycosidases. Thereafter, mucins were solubilized by incubating for 90 min at 37 °C. After centrifugation for 1 min at 15,000 × g, the content of mucins was measured using a fluorimetric assay that discriminates O-linked glycoproteins (mucins) from N-linked glycoproteins (Crowther and Wetmore, 1987). Standard solutions of N-acetylgalactosamine (Sigma-Aldrich, St. Louis, MO, USA) were used to calculate the number of oligosaccharides liberated from mucins during the procedure. The cecal content pH was measured by a pH meter (9618S-10D, Horiba Scientific Ltd., Kyoto, Japan).
Statistical analysis Data are expressed as mean ± SE. Data were analyzed by a Student's t-test. Some data were analyzed using the Spearman rank correlation analysis. For all tests, p < 0.05 was considered statistically significant. Data analysis was performed using Excel Statistics 2016 for Windows (Social Survey Research Information Co., Ltd., Tokyo, Japan).
As shown in Table 2, body weight gain and food intake did not differ between the two groups studied (p > 0.05). Liver weight was unaffected, whereas relative liver weight (%) in the GlcN group was slightly but significantly higher than that in the control group (p < 0.05). Epididymal adipose tissue weight and its relative weight (%) were decreased by GlcN intake (p < 0.05), but no differences were found between the two groups in perirenal adipose tissue weight.
| Control | GlcN1 | |
|---|---|---|
| Initial body wt (g) | 82 ± 1 | 81 ± 1 |
| Final body wt (g) | 419 ± 5 | 408 ± 5 |
| Gains in body wt (g) | 345 ± 5 | 331 ± 5 |
| Food intake (g/56 days) | 1096 ± 6 | 1092 ± 4 |
| Liver wt (g) | 14.3 ± 0.2 | 15.3 ± 0.4 |
| Liver wt (%) | 3.42 ± 0.06 | 3.76 ± 0.07* (+10%) |
| Epididymal adipose tissue wt (g) | 4.08 ± 0.19 | 3.19 ± 0.19* (−22%) |
| Epididymal adipose tissue wt (%) | 0.97 ± 0.05 | 0.76 ± 0.04* (−22%) |
| Perirenal adipose tissue wt (g) | 4.85 ± 0.27 | 4.12 ± 0.35 |
| Perirenal adipose tissue wt (%) | 1.23 ± 0.07 | 1.01 ± 0.08 |
Mean ± SE (n = 8),
Results of the serum parameters determined are presented in Table 3. Parameters associated with liver function, including AST, ALT, LDH, γ-GTP, and albumin were unaffected by GlcN intake (p > 0.05). On the other hand, serum levels of ammonia and ethanol were significantly increased by GlcN intake (+21 % and +12 %, respectively, p < 0.05). Serum levels of urea, uric acid, and creatinine were unaffected by GlcN treatment, whereas serum levels of total cholesterol and LDL cholesterol were significantly increased (+22 % and +38 %, respectively, p < 0.05). Serum glucose and lactic acid levels were unchanged, whereas those of Mg were slightly but significantly increased by GlcN intake (+16 %, p < 0.05).
| Serum parameters | Control | GlcN1 |
|---|---|---|
| AST (Aspartate aminotransferase) (U/L) | 146 ± 6 | 145 ± 9 |
| ALT (Alanine aminotransferase) (U/L) | 37.3 ± 2.5 | 38.6 ± 3.2 |
| LDH (Lactate dehydrogenase) (U/L) | 708 ± 62 | 634 ± 53 |
| γ-GTP (γ-Glutamyl transpeptidase) (U/L) | 0.39 ± 0.12 | 0.44 ± 0.13 |
| Albumin (g/L) | 39.3 ± 0.5 | 42.0 ± 1.8 |
| Ammonia (µmol/L) | 156 ± 6 | 188 ± 9* (+21%) |
| Ethanol (µmol/L) | 1.38± 0.06 | 1.55 ± 0.03*(+12%) |
| Urea (mmol/L) | 5.78 ± 0.18 | 5.97 ± 0.18 |
| Uric acid (µmol/L) | 125 ± 4 | 131 ± 4 |
| Creatinine (µmol/L) | 19.8 ± 0.76 | 18.3 ± 0.76 |
| Triglyceride (mmol/L) | 2.36 ± 0.26 | 2.06 ± 0.23 |
| Total cholesterol (mmol/L) | 2.51 ± 0.10 | 3.06 ± 0.23* (+22%) |
| LDL-cholesterol (mmol/L) | 0.37 ± 0.03 | 0.51 ± 0.05*(+38%) |
| Free fatty acid (µmol/L) | 491 ± 34 | 591± 72 |
| Glucose (mmol/L) | 11.8 ± 0.4 | 12.2 ± 0.6 |
| Lactic acid (mmol/L) | 6.03 ± 0.41 | 5.76 ± 0.77 |
| Ca (µmol/L) | 308 ± 10 | 328 ± 15 |
| Fe (µmol/L) | 5.32 ± 0.18 | 5.01 ± 0.23 |
| Inorganic phosphorus (mmol/L) | 2.95 ± 0.07 | 3.19 ± 0.18 |
| Mg (µmol/L) | 10.9 ± 0.4 | 12.6 ± 0.1*(+16%) |
Mean ± SE (n = 7–8),
We found that GlcN intake markedly increased cecal content weight (+115 %, p < 0.001, Table 4). In addition, dietary GlcN increased the levels of ammonia per gram of contents and per total contents (+27 %, p < 0.05 and +226 %, p < 0.001, respectively), as well as the levels of ethanol per gram contents and per total contents (+93 %, p < 0.05 and +317 %, p < 0.01, respectively). GlcN intake significantly increased the numbers of total bacteria, Clostridium coccoides, and Firmicutes per gram of cecal contents (+79 %, +91 %, and +65 %, respectively, p < 0.05), but did not affect the other bacterial numbers. Expressed per total contents, dietary GlcN significantly increased the numbers of total bacteria (p < 0.001), Lactobacillus (p < 0.05), C. coccoides (p < 0.001), Clostridium leptum (p < 0.001), Bacteroides (p < 0.001), and Firmicutes (p < 0.001), but did not affect those of Bifidobacterium (p > 0.05).
| Cecal parameters | Control | GlcN1 |
|---|---|---|
| Contents wt (g) | 3.14 ± 0.18 | 6.76 ± 0.24*** (+115%) |
| Ammonia (µmol/g contents) | 31.8 ± 1.8 | 40.4 ± 2.5*(+27%) |
| Ammonia (µmol/total contents) | 99 ± 7 | 323 ± 15***(+226%) |
| Ethanol (µmol/g contents) | 1.40 ± 0.28 | 2.70 ± 0.22**(+93%) |
| Ethanol (µmol/total contents) | 4.48 ± 2.93 | 18.69 ± 1.67** (+317%) |
| pH | 7.29 ± 0.05 | 7.21 ± 0.05 |
| Cecal microflora | Copy numbers/g contents | |
| Total bacteria (x1013) | 5.43 ± 1.17 | 9.74 ± 0.27*(+79%) |
| Bifidobacterium spp. (x1011) | 2.20 ± 1.06 | 0.67 ± 0.28 |
| Lactobacillus spp. (x1013) | 2.28 ± 0.43 | 1.91 ± 0.25 |
| Clostridium coccoides (x1013) | 0.97 ± 0.20 | 1.85 ± 0.22*(+91%) |
| Clostridium leptum (x1013) | 3.54 ± 0.96 | 4.47 ± 0.31 |
| Bacteriodes (x1013) | 2.82 ± 0.69 | 4.31 ± 0.40 |
| Firmicutes (x1013) | 2.71 ± 0.49 | 4.46 ± 0.24*(+65%) |
| Cecal microflora | Copy numbers/total contents | |
| Total bacteria (x1013) | 16.7 ± 3.6 | 65.7 ± 2.6***(+293%) |
| Bifidobacterium spp. (x1011) | 6.8 ± 3.2 | 4.7 ± 2.1 |
| Lactobacillus spp. (x1013) | 7.0 ±1.4 | 13.1 ± 1.9*(+87%) |
| Clostridium coccoides (x1013) | 3.0 ± 0.6 | 12.5 ± 1.6***(+317%) |
| Clostridium leptum (x1013) | 10.8 ± 2.9 | 30.4 ± 2.7***(+181%) |
| Bacteriodes (x1013) | 8.5 ± 2.1 | 28.9 ± 2.6***(+240%) |
| Firmicutes (x1013) | 8.3 ± 1.5 | 30.4 ± 2.4***(+266%) |
Mean ± SE (n = 8),
The fecal IgA levels were unaffected as expressed both per gram of dry feces and over 2 days (p > 0.05) (Table 5). Fecal mucin levels over 2 days were significantly increased by GlcN intake (p < 0.05), although those per gram of dry feces were unaffected.
| Fecal parameters | Control | GlcN1 |
|---|---|---|
| Dry wt (g) | 2.71 ± 0.14 | 3.36 ± 0.22 |
| IgA (mg/g fecal dry wt) | 1.55 ± 0.40 | 0.93 ± 0.22 |
| IgA (mg/2 days) | 4.32 ± 1.25 | 3.02 ± 0.61 |
| Mucins (mg/g fecal dry wt) | 3.79 ± 0.51 | 5.27 ± 0.76 |
| Mucins (mg/2 days) | 10.1 ± 1.2 | 17.6 ± 2.9*(+74%) |
Mean ± SE (n = 8),
Cecum ammonia levels were significantly associated with the numbers of total bacteria and C. coccoides (r=0.537 and r=0.501, respectively, p < 0.05), but not with those of Firmicutes (r=0.383, p > 0.05) when expressed per gram of contents. Furthermore, cecal ethanol levels were significantly associated with the numbers of total bacteria and C. coccoides (r=0.541 and r=0.650, respectively, p < 0.05), but not with those of Firmicutes (r=0.432, p > 0.05) per gram of contents. Serum ammonia levels were not associated with the numbers of total bacteria, C. coccoides, and Firmicutes (r=0.426, r=0.345, and r=0,153, respectively, p > 0.05) per gram of contents. Similarly, serum ethanol levels were not associated with the numbers of total bacteria, C. coccoides, and Firmicutes (r=0.426, r=0.375, and r=0.375, respectively, p > 0.05) per gram of contents.
The findings of our study indicated that supplemental GlcN increased the levels of ammonia in the serum and cecum. The high level of circulating ammonia (a risk factor for neuron diseases) is considered to be ascribed to liver dysfunction (impaired urea production from ammonia) and/or to higher production of ammonia by intestinal bacteria. Since our results indicated the absence of any effect of GlcN on the serum parameters related to liver damage (dysfunction), the possibility that GlcN-induced liver dysfunction resulting in higher serum ammonia levels appears to be negated. Further, the results of this study indicated that the increased cecum ammonia levels following GlcN intake were linked to increased numbers of total bacteria and C. coccoides. Clostridium has been suggested to be classified as an ammonia-producing bacterium (Zuo et al., 2017). Therefore, the overgrowth of C. coccoides may cause an elevation in ammonia levels. Our finding concerning the link between bacterial overgrowth and higher ammonia levels appears to be similar to the result reported by Farahmand et al. (2016), which indicated that bacterial overgrowth caused hyperammonemia in a child with Hirschsprung's disease. In addition, it has been reported that D-glucosamine-6-phosphate can be metabolized by glucosamine-6-phosphatase to ammonia and fructose-6-phosphate in Escherichia coli and human tissue (Arreola et al., 2003, Calcagno et al., 1984). Therefore, the possibility that the amino group of GlcN may contribute to increased ammonia levels observed in this study remains. Colon ammonia has been considered to be a risk factor for colon cancer (Fung et al., 2013). However, a previous study indicated that GlcN intake was associated with a reduced risk of colon cancer (Kantor et al., 2016). At present, the cause of this discrepancy is still unknown. Further studies are required to elucidate the pathological implication of increased cecum ammonia by GlcN intake.
This study further showed that supplemental GlcN increased the levels of ethanol in the serum and cecum. High blood ethanol is considered to be associated with alcohol-related diseases, such as alcoholic cirrhosis and alcoholic encephalopathy (Jepsen et al., 2012), heart disease (Criqui and Thomas, 2017), and cancer (Ratna and Mandrekar, 2017). Additionally, high levels of ethanol in the colon have been linked to colon cancer (Tsuruya et al., 2016). Therefore, high GlcN intake may cause an adverse effect on colon health by elevating ethanol. Our findings also indicated that the elevated ethanol levels in the cecum were associated with increased numbers of total bacteria and C. coccoides. However, evidence demonstrating that C. coccoides per se can produce ethanol has not been reported, and this possibility remains to be investigated. Our finding of the link between bacterial overgrowth and increased ethanol levels appears to be similar to the case of small intestinal bacterial overgrowth syndrome linked to the endogenously higher production of ethanol, although the mechanistic link between bacterial overgrowth and increased ethanol production is unknown (Bures et al., 2010).
Our study indicated that supplemental GlcN increased fecal mucins, but did not affect fecal IgA. Mucins perform the function of an intestinal barrier by uniquely protecting and lubricating epithelial surfaces (Wang and Fang, 2003). In the present study, we used fecal mucin levels as an index of intestinal mucin production, as there is a general association between intestinal mucin levels and fecal mucin levels (Ten Bruggencate et al., 2005). However, fecal excretion of mucins depends on various complicated circumstances in the digestive tract, including the microbiota profile. Additionally, mucins consist of various types of mucin proteins, some of which are associated with diseases (Kim and Ho, 2010, Velcich et al., 2002). Therefore, further study is necessary to investigate the profile of several types of mucins in the intestinal contents in order to gain insight into the physiological implication of increased fecal mucins by GlcN intake.
It is noteworthy that supplemental GlcN increased serum levels of total cholesterol and LDL cholesterol. This result was similar to the finding reported by Tannock et al. (2006) that the long-term consumption of GlcN HCl (50 mg/kg body wt/day) for 20 weeks increased plasma cholesterol concentrations in LDL receptor-deficient mice. However, the implication of high total cholesterol and LDL cholesterol following high-dose GlcN intake in relation to atherosclerosis remains to be investigated. Furthermore, in this study, supplemental GlcN slightly but significantly increased the serum levels of Mg. Low circulating Mg levels have been associated with a high risk of various diseases, including type 2 diabetes (Wahid et al., 2017), cardiovascular disease (Lee et al., 2015), hypertension (Han et al., 2017), asthma (Shaikh et al., 2016), and fractures (Kunutsor et al., 2017). However, it is unknown whether the increased serum Mg levels following GlcN intake have suppressive effects on such diseases.
In conclusion, we found that high-dose GlcN increased the levels of ammonia and ethanol in the serum and cecum, which was associated with bacterial overgrowth. This finding implies that high-dose GlcN intake may cause adverse effects by increasing ammonia and ethanol. To the best of our knowledge, this is the first evidence of an increase in the levels of both ammonia and ethanol, which are harmful metabolites, by a dietary factor. The elevated levels of ammonia and ethanol by GlcN intake might have resulted from bacterial overgrowth; however, the possibility remains that the increase in these two metabolites by GlcN contributed to bacterial overgrowth. The intake of high-dose GlcN (4 %) in the diet for 8 weeks was assessed in this investigation; thus, further studies are necessary to elucidate the long-term effect of low-dose dietary GlcN on ammonia and ethanol levels.
Acknowledgments This work was supported by Funds for the Development of Human Resources in Science and Technology under the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan through the Home for Innovative Researchers and Academic Knowledge Users (HIRAKU) consortium, Hiroshima University, Japan.
The authors declare that they have no conflicts of interest.
alanine aminotransferase
ASTaspartate aminotransferase
CHDcoronary heart disease
CRCcolorectal cancer
CVDcoronary vascular disease
ELISAenzyme-linked immunosorbent assay
γ-GTPgamma-glutamyl transpeptidase
GlcND-glucosamine HCl
HPFSHealth Professional follow-up study
HEhepatic encephalopathy
IgAimmunoglobulin
LDHlactate dehydrogenase
LDLlow-density-lipoprotein
Mgmagnesium
NHSNurses' Health Study
qPCRreal-time quantitative PCR
SEstandard error
