Fatty liver and serum cholinesterase are independently correlated with HbA1c levels: Cross-sectional analysis of 5384 people

Abstract

Objectives:

To examine the association between glycosylated haemoglobin (HbA1c) and fatty liver markers.

Methods:

This cross-sectional analysis stratified subjects into quintiles based on HbA1c. Fatty liver using ultrasonography scores (FLUS) were assigned as follows: 2 points, moderate or severe fatty liver; 1 point, mild fatty liver; and 0 points, normal liver. Subjects with viral hepatitis, alcohol intake >175 g/week or receiving hypoglycaemic treatment were excluded.

Results:

The study included 5384 subjects. Serum cholinesterase (ChE) and FLUS showed a significant graded increase with increasing HbA1c. In linear regression analysis stratified by body mass index (BMI) and age, ChE and FLUS were significantly associated with lower (1 + 2) and higher (3 + 4 + 5) HbA1c quintiles, respectively, independent of BMI and age.

Conclusions:

The findings show that both ChE and FLUS are significantly correlated with HbA1c, independent of BMI and age.

Keywords

Cholinesterase non-alcoholic fatty liver disease HbA1c

Introduction

The metabolic syndrome is known to cause atherosclerosis, and is defined as central obesity (indicated by waist circumference) plus any two of the following factors: elevated triglycerides, reduced high-density-lipoprotein cholesterol (HDL-C), elevated blood pressure and elevated fasting plasma glucose.¹ The metabolic syndrome is associated with non-alcoholic fatty liver disease (NAFLD)² and is correlated with serum alanine aminotransferase (ALT) concentration in a dose-dependent manner (within the reference range),³ suggesting that NAFLD and ALT concentration can be used in tandem to identify at-risk subjects.⁴ Elevated aminotransferases are also independently associated with the metabolic syndrome in patients with diabetes who have no ultrasonographic evidence of fatty liver.⁵ Since insulin resistance is a key mechanism of NAFLD, the metabolic syndrome and glycaemic disorder,⁶ NAFLD is also independently associated with HbA1c concentration in non-diabetic individuals.⁷ The FIN-D2D survey reported in 2010 that the prevalence of the metabolic syndrome and type 2 diabetes are significantly increased in subjects with NAFLD compared with those with normal liver function tests.⁸

It is likely that NAFLD is associated with both the metabolic syndrome and glycaemic disorder (as indicated by HbA1c), but ultrasonographic findings and ALT concentrations may not be definitive markers of fatty liver in all individuals. Serum cholinesterase (ChE) concentration is increased in patients with fatty liver, but few studies have investigated the direct association between glycaemic disorder and ChE and other markers of fatty liver. A single, small-scale study found that ChE concentrations were higher in patients with non-alcoholic steatohepatitis (NASH), which was also associated with diabetes.⁹

The aim of this preliminary study, therefore, was to examine the hypothesis that fatty liver (diagnosed via ultrasonography) and ChE are associated with HbA1c.

Subject and methods

Study population

This cross-sectional retrospective analysis included workers or retirees aged between 19 and 90 years who were undergoing an annual medical check-up as part of the healthcare system at Jikei University Hospital Health Care Centre, Tokyo, Japan between January 2006 and December 2006. Exclusion criteria were: (i) treatment with hypoglycaemic agents; (ii) hepatitis B surface antigen positive; (iii) hepatitis C antibody positive; and (iv) ethanol intake >175 g/week.

All participants provided written informed consent prior to enrolment in the study. The study was approved by the Institutional Review Board of Jikei University School of Medicine and carried out in accordance with the Declaration of Helsinki.

Ultrasonography

Experienced technicians, who were unaware of the aims of the study and blinded to laboratory findings, evaluated fatty liver via abdominal ultrasonography. All medical sonographers were registered with the Japan Society of Ultrasonics in Medicine and had been trained at Jikei University Hospital. Study-specific fatty liver using ultrasonography scores (FLUS) were assigned as follows: 2 points, moderate or severe fatty liver (deep attenuation, vascular blurring, a fatty bandless sign and/or a brighter liver compared with the spleen^10–12); 1 point, mild fatty liver (bright liver, focal fatty sparing¹⁰ and/or high hepato–renal echo contrast^11,13); and 0 points, normal liver.

Data collection

Medical history and lifestyle information including alcohol consumption and smoking were documented by a professional nurse. Subject height and weight were recorded with subjects dressed in lightweight indoor clothes, and body mass index (BMI) calculated as weight/height² (kg/m²). Waist circumference while undressed and standing was measured (to the nearest cm) using a nonelastic soft tape at the umbilical level when the navel was between the lowest rib and the level of the iliac crest, or midway between the lowest rib and the level of the iliac crest. Blood pressure (BP) was measured using a standard mercury manometer after 5 min rest in a sitting position. Following the blood pressure measurement, venous blood specimens (12 ml for serum and 1.8 ml for plasma) were drawn between 09:00 and 10:30 hours, after an overnight fast. Plasma and serum were stored at 10–15℃ prior to analysis within 4 h.

Laboratory analyses

Fasting plasma glucose (FPG) concentrations were determined using the glucose oxidase immobilized oxygen electrode method (GA08III; A&T Corporation, Kanagawa, Japan).¹⁴ HbA1c (NGSP equivalent value)¹⁵ was determined with high-performance liquid chromatography (HLC-723G9; Tosoh Corporation, Tokyo, Japan). Enzymatic methods were used to analyse serum high-density-lipoprotein cholesterol (HDL-C) (MetaboLead HDL-C; Kyowa Medex, Tokyo, Japan), low-density-lipoprotein cholesterol (LDL-C) and triglyceride (TG) (Cholestest LDL and Cholestest TG, respectively; Sekisui Medical, Tokyo, Japan), and aspartate aminotransferase (AST) and alanine aminotransferase (ALT) (AST-J2 and ALT-J2, respectively; Wako Pure Chemical Industries, Osaka, Japan). Serum ChE and γ-glutamyltransferase (GGT) levels were determined using a butyrylthiocholine iodide method and γ-glutamyl-3-carboxy-4-nitroanilide method (both Wako Pure Chemical Industries, Ltd, Osaka, Japan), respectively. Serum hepatitis B virus surface antigen and hepatitis C virus antibody (3rd generation) were quantified by chemiluminescent enzyme immunoassay (CLEIA) (Architect i2000SR; Abbott Japan, Tokyo, Japan). Serum C-reactive protein (CRP) and uric acid were quantified by latex agglutination immunoassay (Iatro CRP-EX, Mitsubishi Chemical Medicine Corporation, Tokyo, Japan) and uricase peroxidase method (Wako Pure Chemical Industries), respectively. All samples were analysed in the same laboratory at Jikei University Hospital.

Statistical analyses

Subjects were stratified into sex-specific quintile groups by HbA1c concentration (groups 1/5 to 5/5). The same HbA1c cut-off points are applied to both men and women in clinical practice, but subjects were stratified by sex in the present study in order to permit analyses of possible sex-specific differences. Participant characteristics were presented as mean ± SD for continuous variables and proportions (%) for dichotomous variables. The association between HbA1c and all other variables was evaluated as a trend-test with linear regression models for continuous variables and logistic regression models for dichotomous variables. In addition, the numbers of individuals evaluated for ChE and FLUS were tested with the Jonckheere–Terpstra test. Multiple regression models stratified by BMI and age were further used to evaluate determinants of HbA1c. Subjects were stratified according to median BMI and age for the linear regression analysis. The analysis was also performed with stratifications for HbA1c quintiles in order to test for differences in associations between those with lower and those with higher quintiles.

Stata statistical package version 10.1 (StataCorp LP, College Station, Texas, USA) was used for statistical analysis. P-values < 0.05 were considered statistically significant.

Results

The study recruited 7984 subjects (5437 male/2547 female; mean age, 48.8 years; age range, 19–90 years). A total of 2600 subjects were excluded (receiving hypoglycaemic agents n = 248; alcohol consumption >175 g/week n = 2216; hepatitis B positive n = 51; hepatitis C positive n = 29; missing HbA1c data n = 56). The final analysis included 5384 subjects (3173 male/2211 female; mean age, 48.0 years; age range, 19–90 years). A total of 5167 subjects underwent ultrasonography and were assigned FLUS scores. Serum ChE levels were determined in 3973 subjects. Demographic and clinical characteristics of the subjects stratified by sex and HbA1c quintile are given in Tables 1a and 1b.

Table 1a.

Demographic and clinical characteristics of male subjects included in a cross-sectional analysis of the association between glycosylated haemoglobin (HbA1c) and markers of fatty liver, stratified according to HbA1c quintile.

Parameter	Men
	HbA1c quintile					Total	Statistical significance of trend
	1	2	3	4	5	Total	Statistical significance of trend
n	863 (27.2)	726 (22.9)	361 (11.4)	705 (22.2)	518 (16.3)	3173 (100.0)	NS
Age, years	44 ± 10	47 ± 11	49 ± 11	52 ± 11	56 ± 10	49 ± 11	P < 0.001^a
Waist circumference, cm	82 ± 9	83 ± 9	84 ± 9	85 ± 9	87 ± 10	84 ± 9	P < 0.001^a
BMI, kg/m²	23.0 ± 2.7	23.2 ± 2.7	23.7 ± 3.1	23.8 ± 3.2	24.5 ± 3.2	23.5 ± 3.0	P < 0.001^a
HbA1c, %	5.1 ± 0.2	5.5 ± 0.1	5.6 ± 0.0	5.8 ± 0.1	6.5 ± 0.9	5.6 ± 0.6	NA
FPG, mg/dl	94 ± 7	96 ± 7	98 ± 8	100 ± 8	116 ± 27	100 ± 15	P < 0.001^a
TG, mg/dl	111 ± 87	110 ± 60	120 ± 76	126 ± 70	145 ± 100	121 ± 80	P < 0.001^a
HDL-C, mg/dl	57 ± 13	57 ± 13	55 ± 13	54 ± 12	53 ± 13	55 ± 13	P < 0.001^a
Systolic BP, mmHg	116 ± 13	116 ± 14	118 ± 13	119 ± 14	121 ± 14	118 ± 13	P < 0.001^a
Diastolic BP, mmHg	73 ± 9	73 ± 10	74 ± 9	75 ± 10	76 ± 10	74 ± 10	P < 0.001^a
AST, U/l	22 ± 12	22 ± 9	23 ± 10	24 ± 10	25 ± 12	23 ± 11	P < 0.001^a
ALT, U/l	25 ± 15	26 ± 19	26 ± 18	28 ± 21	31 ± 21	27 ± 19	P < 0.001^a
GGT, U/l	44 ± 55	41 ± 41	45 ± 44	46 ± 47	59 ± 75	46 ± 53	P < 0.001^a
ChE evaluations, n	616 (26.5)	538 (23.2)	266 (11.5)	524 (22.6)	379 (16.3)	2323 (100.0)	NS^b
ChE, mU/dl	5562 ± 1018	5703 ± 1006	5771 ± 935	5736 ± 1024	6014 ± 1082	5732 ± 1028	P < 0.001^a
FLUS evaluations 0/1/2 points, n	568/141/108	455/109/126	194/65/84	349/125/197	197/99/202	1763/539/717	NS/NS/NS^b
FLUS 1–2 points	249 (30.5)	235 (34.1)	149 (43.4)	322 (48.0)	301 (60.4)	1256 (41.6)	P < 0.001^a
FLUS 2 points	108 (13.2)	126 (18.3)	84 (24.5)	197 (29.4)	202 (40.6)	717 (23.7)	P < 0.001^a

Data are n (%) or mean ± SD.

Linear regression model for continuous variables and logistic regression model for dichotomous variables, treating HbA1c quintile groups as continuous predictor.

Jonckheere–Terpstra test.

NS, not statistically significant (P ≥ 0.05); BMI, body mass index; NA, not applicable; FPG, fasting plasma glucose; TG, triglyceride; HDL-C, high-density-lipoprotein cholesterol; BP, blood pressure; AST, aspartate aminotransferase; ALT, alanine aminotransferase; GGT, γ-glutamyltransferase; ChE, cholinesterase; FLUS, fatty liver using ultrasonography scores.

Table 1b.

Demographic and clinical characteristics of female subjects included in a cross-sectional analysis of the association between glycosylated haemoglobin (HbA1c) and markers of fatty liver, stratified according to HbA1c quintile.

Parameter	Women
	HbA1c quintile					Total	Statistical significance of trend
	1	2	3	4	5	Total	Statistical significance of trend
n	611 (27.6)	521 (23.6)	245 (11.1)	404 (18.3)	430 (19.4)	2211 (100.0)	NS
Age, years	41 ± 8	44 ± 10	46 ± 11	50 ± 11	55 ± 10	46 ± 11	P < 0.001^a
Waist circumference, cm	73 ± 8	74 ± 8	75 ± 8	77 ± 9	80 ± 11	76 ± 9	P < 0.001^a
BMI, kg/m²	20.5 ± 2.7	20.7 ± 2.8	20.8 ± 2.9	21.3 ± 3.0	22.4 ± 3.8	21.1 ± 3.1	P < 0.001^a
HbA1c, %	5.1 ± 0.2	5.4 ± 0.1	5.5 ± 0.0	5.6 ± 0.1	6.0 ± 0.4	5.5 ± 0.4	NA
FPG, mg/dl	90 ± 6	91 ± 7	94 ± 7	94 ± 7	101 ± 15	93 ± 10	P < 0.001^a
TG, mg/dl	66 ± 32	71 ± 36	75 ± 57	81 ± 41	95 ± 49	77 ± 43	P < 0.001^a
HDL-C, mg/dl	71 ± 13	73 ± 15	72 ± 17	70 ± 16	67 ± 15	71 ± 15	P < 0.001^a
SBP, mmHg	106 ± 13	108 ± 14	109 ± 15	111 ± 15	116 ± 16	110 ± 15	P < 0.001^a
DBP, mmHg	66 ± 9	67 ± 9	67 ± 9	69 ± 9	72 ± 10	68 ± 10	P < 0.001^a
AST, U/l	18 ± 6	19 ± 5	20 ± 7	21 ± 9	22 ± 10	20 ± 7	P < 0.001^a
ALT, U/l	15 ± 7	15 ± 7	17 ± 10	18 ± 11	21 ± 15	17 ± 11	P < 0.001^a
GGT, U/l	20 ± 24	20 ± 13	24 ± 34	22 ± 18	30 ± 27	23 ± 23	P < 0.001^a
ChE evaluations, n	430 (26.1)	367 (22.2)	180 (10.9)	307 (18.6)	366 (22.2)	1650 (100.0)	NS^b
ChE, mU/dl	4598 ± 875	4949 ± 992	4905 ± 961	5288 ± 1349	5627 ± 1125	5066 ± 1130	P < 0.001^a
FLUS evaluations 0/1/2 points, n	546/30/9	474/21/12	211/22/8	332/37/20	293/66/67	1856/176/116	NS/NS/NS^b
FLUS 1–2 points, %	39 (6.7)	33 (6.5)	30 (12.4)	57 (14.7)	133 (31.2)	292 (13.6)	P < 0.001^a
FLUS 2 points, %	9 (1.5)	12 (2.4)	8 (3.3)	20 (5.1)	67 (15.7)	116 (5.4)	P < 0.001^a

Data are n (%) or mean ± SD.

Linear regression model for continuous variables and logistic regression model for dichotomous variables, treating HbA1c quintile groups as continuous predictor.

Jonckheere–Terpstra test.

In both men and women, there was a significant graded increase in BMI, mean age and FPG from quintile 1 to quintile 5 (P < 0.001 for each trend; Tables 1a and 1b).

Data regarding the metabolic syndrome and hepatic function markers are shown in Tables 1a and 1b. Waist circumference, TG, and systolic and diastolic BP showed a significant graded increase, and HDL-C was significantly decreased from quintile 1 to quintile 5 in both men and women (P < 0.001 for each trend). In addition, there were significant graded increases in AST and ALT from quintile 1 to quintile 5 in both men and women (P < 0.001 for each trend).

Linear regression analysis of subjects stratified according to median BMI (22.3 kg/m²) indicated that FLUS score was significantly associated with HbA1c independent of BMI in the total study population (P < 0.001 for each association; Tables 2a and 2b). The association in quintiles 3 + 4 + 5 combined (β = 0.138, P < 0.001 and β = 0.144, P < 0.001 for BMI ≤ 22.3 kg/m² and >22.3 kg/m², respectively; Tables 2a and 2b) contributed to these findings. In addition, ChE was also significantly associated with HbA1c independent of BMI in the total study population (P < 0.001 for each association; Tables 2a and 2b). In subgroup analysis, the association in quintiles 1 + 2 combined (β = 0.126, P < 0.001; Table 2a) contributed to the findings in the total study population, as well as in subjects with BMI >22.3 kg/m² (Table 2b).

Table 2a.

Standardized linear regression analysis of determinants of glycosylated haemoglobin (HbA1c) in subjects with BMI ≤ 22.3 kg/m², in the total study population and when stratified according to HbA1c quintile.

Parameter	All subjects n = 2712		Quintiles 1 + 2 n = 1542		Quintiles 3 + 4 + 5 n = 1170
Parameter	β	Statistical significance	β	Statistical significance	β	Statistical significance
Age, years	0.359	P < 0.001	0.120	P < 0.001	0.290	P < 0.001
FLUS, points	0.083	P < 0.001	−0.039	NS	0.138	P < 0.001
ChE, mU/dl	0.089	P < 0.001	0.126	P < 0.001	0.051	NS
Diastolic BP, mmHg	−0.022	NS	0.058	NS	−0.037	NS
Systolic BP, mmHg	−0.012	NS	−0.024	NS	−0.007	NS
BMI, kg/m²	0.010	NS	0.046	NS	0.054	NS
AST, U/l	−0.108	P = 0.002	−0.021	NS	−0.165	P = 0.005
ALT, U/l	0.077	P = 0.021	0.028	NS	0.098	NS
Waist circumference, cm	−0.071	P = 0.011	−0.043	NS	−0.124	P = 0.004
LDL-C, mg/dl	0.097	P < 0.001	0.127	P < 0.001	0.023	NS
GGT, U/l	0.040	NS	−0.068	NS	0.145	P < 0.001
Serum uric acid, mg/dl	−0.026	NS	0.083	P = 0.014	0.048	NS
CRP, mg/dl	0.015	NS	0.004	NS	0.009	NS
TG, mg/dl	0.028	NS	0.002	NS	−0.022	NS

FLUS, fatty liver using ultrasonography scores; NS, not statistically significant (P ≥ 0.05); ChE, cholinesterase; BP, blood pressure; BMI, body mass index; AST, aspartate aminotransferase; ALT, alanine aminotransferase; LDL-C, low-density-lipoprotein cholesterol; GGT, γ-glutamyltransferase; CRP, C-reactive protein; TG, triglyceride.

Results of linear regression analysis of subjects stratified according to median age (46 years) are given in Tables 3a and 3b. HbA1c was significantly associated with FLUS in all subjects (P < 0.001 for each association; Tables 3a and 3b). The association in quintiles 3 + 4 + 5 combined (β = 0.258, P < 0.001 and β = 0.114, P = 0.001 for age ≤46 and >46 years, respectively; Tables 3a and 3b) contributed to these findings. HbA1c was also significantly associated with ChE in all subjects (P < 0.05; Tables 3a and 3b). The association in quintiles 1 + 2 combined (β = 0.107, P = 0.003 and β = 0.149, P = 0.001 for age ≤46 and >46 years, respectively; Tables 3a and 3b) contributed to these findings.

Table 2b.

Standardized linear regression analysis of determinants of glycosylated haemoglobin (HbA1c) in subjects with BMI > 22.3 kg/m², in the total study population and when stratified according to HbA1c quintile.

Parameter	All subjects n = 2672		Quintiles 1 + 2 n = 1179		Quintiles 3 + 4 + 5 n = 1493
Parameter	β	Statistical significance	β	Statistical significance	β	Statistical significance
Age, years	0.270	P < 0.001	0.204	P < 0.001	0.121	P < 0.001
FLUS, points	0.179	P < 0.001	0.015	NS	0.144	P < 0.001
ChE, mU/dl	0.110	P < 0.001	0.099	P = 0.013	0.090	P = 0.006
Diastolic BP, mmHg	–0.115	P = 0.004	0.041	NS	−0.135	P = 0.013
Systolic BP, mmHg	0.099	P = 0.012	−0.062	NS	0.152	P = 0.006
BMI, kg/m²	0.086	P = 0.008	0.032	NS	0.050	NS
AST, U/l	−0.050	NS	0.062	NS	−0.095	NS
ALT, U/l	0.050	NS	−0.003	NS	0.094	NS
Waist circumference, cm	−0.039	NS	−0.051	NS	−0.022	NS
LDL-C, mg/dl	0.037	NS	0.114	P = 0.001	−0.040	NS
GGT, U/l	0.061	P = 0.015	−0.025	NS	0.119	P = 0.001
Serum uric acid, mg/dl	−0.060	P = 0.007	0.064	NS	−0.091	P = 0.003
CRP, mg/dl	0.031	NS	−0.033	NS	0.009	NS
TG, mg/dl	0.007	NS	−0.008	NS	0.010	NS

Table 3a.

Standardized linear regression analysis of determinants of glycosylated haemoglobin (HbA1c) in subjects aged ≤46 years, in the total study population and when stratified according to HbA1c quintile.

Parameter	All subjects n = 2700		Quintiles 1 + 2 n = 1794		Quintiles 3 + 4 + 5 n = 906
Parameter	β	Statistical significance	β	Statistical significance	β	Statistical significance
Age, years	0.100	P < 0.001	0.068	P = 0.016	0.084	P = 0.040
FLUS, points	0.234	P < 0.001	−0.030	NS	0.258	P < 0.001
ChE, mU/dl	0.074	P = 0.013	0.107	P = 0.003	0.077	NS
Diastolic BP, mmHg	−0.028	NS	0.045	NS	−0.095	NS
Systolic BP, mmHg	0.012	NS	−0.053	NS	0.065	NS
BMI, kg/m²	0.100	P = 0.038	0.027	NS	0.147	NS
AST, U/l	−0.026	NS	−0.014	NS	−0.124	NS
ALT, U/l	−0.016	NS	0.077	NS	0.057	NS
Waist circumference, cm	−0.088	NS	−0.080	NS	−0.107	NS
LDL-C, mg/dl	0.096	P < 0.001	0.140	P < 0.001	0.033	NS
GGT, U/l	0.059	P = 0.046	−0.040	NS	0.074	NS
Serum uric acid, mg/dl	−0.010	NS	0.086	P = 0.013	−0.048	NS
CRP, mg/dl	0.025	NS	−0.003	NS	0.045	NS
TG, mg/dl	−0.069	P = 0.017	−0.035	NS	−0.109	P = 0.022

Table 3b.

Standardized linear regression analysis of determinants of glycosylated haemoglobin (HbA1c) in subjects aged > 46 years, in the total study population and when stratified according to HbA1c quintile.

Parameter	All Subjects n = 2684		Quintiles 1 + 2 n = 927		Quintiles 3 + 4 + 5 n = 1757
Parameter	β	Statistical significance	β	Statistical significance	β	Statistical significance
Age, years	0.144	P < 0.001	0.095	P = 0.018	0.081	P = 0.004
FLUS, points	0.136	P < 0.001	−0.002	NS	0.114	P = 0.001
ChE, mU/dl	0.115	P < 0.001	0.149	P = 0.001	0.065	P = 0.022
Diastolic BP, mmHg	−0.144	P < 0.001	0.047	NS	−0.131	P = 0.009
Systolic BP, mmHg	0.111	P = 0.007	−0.020	NS	0.135	P = 0.008
BMI, kg/m²	0.078	P = 0.034	0.065	NS	0.095	NS
AST, U/l	−0.083	P = 0.025	0.074	NS	−0.138	P = 0.002
ALT, U/l	0.131	P = 0.001	−0.044	NS	0.168	P = 0.001
Waist circumference, cm	−0.063	NS	−0.055	NS	−0.083	NS
LDL-C, mg/dl	0.037	NS	0.114	P = 0.003	−0.048	NS
GGT, U/l	0.047	NS	−0.067	NS	0.142	P < 0.001
Serum uric acid, mg/dl	−0.042	NS	0.077	NS	−0.041	NS
CRP, mg/dl	0.024	NS	−0.045	NS	−0.020	NS
TG, mg/dl	0.070	P = 0.004	0.021	NS	0.080	P = 0.007

Sequential changes in the value of β suggested that ChE was correlated more strongly with HbA1c than FLUS in lower HbA1c quintiles (1 + 2), while FLUS was more important than ChE in higher quintiles (3 + 4 + 5).

Discussion

Since the establishment of the concept of NAFLD and non-alcoholic steatohepatitis (NASH), the role of fatty liver in metabolic disease has become of interest. Fatty liver markers within or slightly above the upper limit of the normal range have been shown to be pathogenic, and fatty metamorphosis of hepatocytes leads to insulin resistance,^16–18 with optimum levels of metabolic markers being below mean values in healthy subjects in modern developed countries.^7,19–22 Fatty liver markers are not included among the criteria of the metabolic syndrome, however. We therefore compared the criteria of the metabolic syndrome with fatty liver markers in terms of glycaemic status.

There is a lack of consensus regarding the ideal marker for glycaemic status, particularly in the postprandial state. Although elevated FPG increases medical care costs in patients with impaired fasting glucose,²³ post-challenge hyperglycaemia is associated with death and macrovascular complications even before the development of overt diabetes.²⁴ Slightly elevated HbA1c is useful as it suggests a history of slightly raised plasma glucose, especially in the postprandial state,²⁵ and this has been found to be significantly associated with increased mortality even in people without diabetes.^26–28 Although HbA1c is an internationally recognised marker of glycaemic status,²⁹ it is not included among the criteria for the metabolic syndrome.

Plasma ChE is a pseudocholinesterase found primarily in the liver, and it is increased in patients with fatty liver, nephrotic syndrome, Grave’s disease and hepatocellular carcinoma.^30–33 In the present study, only two individuals had nephrotic syndrome and their exclusion did not influence the results. Patients with overt and untreated Grave’s disease and hepatocellular carcinoma were not included in the annual health check-up, and were therefore not enrolled in the study cohort. ChE is known to decrease in patients with hypothyroidism, liver cirrhosis, fulminant hepatitis or sepsis, and those undergoing kidney dialysis or who are pregnant.^32,34 Patients with these conditions were also not included in the annual health check-up programme. Elevated ChE was associated with HbA1c in subjects in quintiles 1 + 2 in the present study. This may lead to worsening of HbA1c status, as indicated by the association between HbA1c and FLUS points in quintiles 3 + 4 + 5.

It has been suggested that biochemical fatty liver markers be included among the criteria of the metabolic syndrome,³⁵ but previous fatty liver indices did not include ChE.^7,36 Elevated ChE was significantly associated with raised HbA1c in the present study.

Data from the present study indicate an increase in ALT from quintile 1 to quintile 5. ALT is used in the scoring system during selection of patients for liver biopsy,³⁷ and for scoring NASH and the risk of incident diabetes.^36,38 Elevated ALT can be used to predict the onset of type 2 diabetes and the metabolic syndrome even within the upper limit of the normal range,^19,20 and increased ALT within the normal range is associated with cardiovascular risk factors independent of obesity.²¹ A Korean study found a direct association between aminotransferase concentration (even within the normal range) and mortality from liver disease.²² Nevertheless, ChE and FLUS were shown to be better markers of fatty liver than ALT in our present study.

Overt NAFLD has been identified as a risk factor for type 2 diabetes in Japanese subjects.³⁹ Magnetic resonance imaging is the most reliable noninvasive imaging method for assessing intrahepatic content,^40,41 but ultrasonography is the most commonly available resource and gives a high degree of certainty during diagnosis of fatty liver in large cohort studies.^13,42 Quantitative techniques aimed at evaluation of fatty liver by ultrasonography have yet to be established. Ultrasonography has been used to assess the presence or absence of fatty liver and to determine the proportion of subjects with NAFLD.⁷ There is a positive correlation between the degree of fatty changes and an increase in echogenicity,¹¹ however, and we therefore attempted to evaluate fatty liver quantitatively via FLUS. A bright liver, focal fatty sparing or high hepato–renal echo contrast indicates mild fatty liver, as described.^10,11,13 Since fatty changes <10–20% cannot be reliably detected by ultrasonography,¹¹ ChE may be a useful marker in these subjects. Fatty liver diagnosed with ultrasonography (FLUS) was significantly associated with raised HbA1c in higher quintiles in the present study.

A slight elevation in ChE level was associated with glycaemic status as indicated by lower HbA1c in the current study (quintiles 1 + 2), whereas FLUS scores were important in higher quintiles (3 + 4 + 5 combined). This result is in line with the findings of others,⁵ and may explain why there is no study reporting that fatty liver detected by ultrasonography is generally associated with the full spectrum of HbA1c.

The present study is limited by the fact that the data were not sufficient to provide cause–effect analysis due to its cross-sectional nature. It is not possible to determine whether higher HbA1c leads to more severe fatty liver/ChE or whether higher fatty liver scores/ChE result in higher HbA1c. Longitudinal studies are required to clarify this issue. The present study did not include specific details of antihypertensive drug treatment. In addition, patients with undiagnosed or untreated diabetes may have been included in our analysis as we did not perform oral glucose tolerance testing. Although smoking is a known risk factor for diabetes,⁴³ smoking (self-reported information) was not significantly associated with HbA1c in the present study. In addition, FLUS and ChE data were not available for all participants. It is not possible, therefore, to exclude the possibility that these issues might impact on our findings.

In conclusion, our study shows that a slight elevation in plasma ChE is correlated with lower HbA1c quintile, while FLUS is important thereafter (independent of BMI and age). A change in marker from ChE to FLUS was observed with increasing HbA1c. Both ChE and FLUS are significantly correlated with HbA1c level and are therefore useful markers, although their contribution may be at different ends of the HbA1c distribution. These markers may enable early intervention, which is known to have a beneficial effect on the improvement of fatty liver.⁴⁴

Declaration of conflicting interest

The authors declare that there are no conflicts of interest.

Funding

Data analysis, writing of the manuscript and publication were supported by Japanese Grants-in-Aid for Scientific Research (KAKENHI 22590609) by the Ministry of Education, Culture, Sports, Science and Technology, Japan.

Footnotes

Acknowledgments

We are grateful to Naoko Tajima, Professor Emerita (Jikei University School of Medicine) for helping in this study.

References

International Diabetes Federation. IDF consensus worldwide definition of the metabolic syndrome, http://www.idf.org/webdata/docs/IDF_Meta_def_final.pdf (accessed 29 January 2014).

Shi

Liu

. Non-alcoholic fatty liver disease: an early mediator predicting metabolic syndrome in obese children? World J Gastroenterol 2011; 17: 735–742.

Kim

Sohn

. Association of serum alanine aminotransferase and γ-glutamyltransferase levels within the reference range with metabolic syndrome and nonalcoholic fatty liver disease. Korean J Hepatol 2011; 17: 27–36.

Hou

Zhu

. Non-alcoholic fatty liver disease’s prevalence and impact on alanine aminotransferase associated with metabolic syndrome in the Chinese. J Gastroenterol Hepatol 2011; 26: 722–730.

Esteghamati

Jamali

Khalilzadeh

. Metabolic syndrome is linked to a mild elevation in liver aminotransferases in diabetic patients with undetectable non-alcoholic fatty liver disease by ultrasound. Diabetol Metab Syndr 2010; 2: 65–65.

Jacobs

van Greevenbroek

van der Kallen

. The association between the metabolic syndrome and alanine amino transferase is mediated by insulin resistance via related metabolic intermediates (the Cohort on Diabetes and Atherosclerosis Maastricht [CODAM] study). Metabolism 2011; 60: 969–975.

Bae

Cho

Lee

. Impact of nonalcoholic fatty liver disease on insulin resistance in relation to HbA1c levels in nondiabetic subjects. Am J Gastroenterol 2010; 105: 2389–2395.

Kotronen

Yki-Järvinen

Männistö

. Non-alcoholic and alcoholic fatty liver disease – two diseases of affluence associated with the metabolic syndrome and type 2 diabetes: the FIN-D2D survey. BMC Public Health 2010; 10: 237–237.

Kojima

Sakurai

Uemura

. Difference and similarity between non-alcoholic steatohepatitis and alcoholic liver disease. Alcohol Clin Exp Res 2005; 29: 259S–263S.

10.

Eisenberg

. Clinical imaging: an atlas of differential diagnosis, 4th edn. Philadelphia, PA: Lippincott Williams & Wilkins, 2003, pp. 513–513.

11.

Kunz

Kuntz

. Hepatology: principles and practice, Heidelberg: Springer-Verlag Berlin, 2002, pp. p.117, 527–p.117, 527.

12.

Boyer

Manns

Sanyal

. Zakin and Boyer’s hepatology, Philadelphia, PA: Elsevier Saunders, 2012, pp. 960–960.

13.

Osawa

Mori

. Sonographic diagnosis of fatty liver using a histogram technique that compares liver and renal cortical echo amplitudes. J Clin Ultrasound 1996; 24: 25–29.

14.

Dimcheva

Horozova

Jordanova

. A glucose oxidase immobilized electrode based on modified graphite. Z Naturforsch C 2002; 57: 705–711.

15.

The Committee of Japan Diabetes Society on the diagnostic criteria of diabetes mellitus. Report of the Committee on the classification and diagnostic criteria of diabetes mellitus. J Jpn Diabetes Soc 2010; 53: 450–467.

16.

Cai

Yuan

Frantz

. Local and systemic insulin resistance resulting from hepatic activation of IKK-beta and NF-kappa B. Nat Med 2005; 11: 183–190.

17.

Sargin

Uygur-Bayramiçli

Sargin

. Asociation of nonalcoholic fatty liver disease with insulin resistance: is OGTT indicated in non-alcoholic fatty liver disease? J Clin Gastroenterol 2003; 37: 399–402.

18.

Parekh

Anania

. Abnormal lipid and glucose metabolism in obesity: implications for non-alcoholic fatty liver disease. Gastroenterology 2007; 132: 2191–2207.

19.

Hanley

Williams

Festa

. Liver markers and development of the metabolic syndrome: the insulin resistance atherosclerosis study. Diabetes 2005; 54: 3140–3147.

20.

Ford

Schulze

Bergmann

. Liver enzymes and incident diabetes: findings from the European Prospective Investigation into Cancer and Nutrition (EPIC)-Potsdam Study. Diabetes Care 2008; 31: 1138–1143.

21.

Kong

Choi

Cockram

. Independent associations of alanine aminotransferase (ALT) levels with cardiovascular risk factor clustering in Chinese adolescents. J Hepatol 2008; 49: 115–122.

22.

Kim

Nam

Jee

. Normal serum aminotransferase concentration and risk of mortality from liver diseases: prospective cohort study. BMJ 2004; 328: 983–983.

23.

Nichols

Brown

. Higher medical care costs accompany impaired fasting glucose. Diabetes Care 2005; 28: 2223–2229.

24.

The DECODE Study Group. Glucose tolerance and mortality: comparison of WHO and American Diabetes Association diagnostic criteria. The DECODE study group. European Diabetes Epidemiology Group. Diabetes Epidemiology: Collaborative analysis Of Diagnostic criteria in Europe. Lancet 1999; 354: 617–621.

25.

Monnier

Lapinski

Colette

. Contributions of fasting and postprandial plasma glucose increments to the overall diurnal hyperglycemia of type 2 diabetic patients: variations with increasing levels of HbA(1c). Diabetes Care 2003; 26: 881–885.

26.

Brewer

Wright

Travier

. A New Zealand linkage study examining the associations between A1C concentration and mortality. Diabetes Care 2008; 31: 1144–1149.

27.

Santos-Oliveira

Purdy

da Silva

. Haemoglobin A1c levels and subsequent cardiovascular disease in persons without diabetes: a meta-analysis of prospective cohorts. Diabetologia 2011; 54: 1327–1334.

28.

Selvin

Steffes

Zhu

. Glycated hemoglobin, diabetes, and cardiovascular risk in nondiabetic adults. N Engl J Med 2010; 362: 800–811.

29.

NGSP Steering Committee. NGSP Protocol, http://www.ngsp.org/protocol.asp (accessed 29 January 2014).

30.

Nomura

Ohnishi

Koen

. Serum cholinesterase in patients with fatty liver. J Clin Gastroenterolol 1986; 8: 599–602.

31.

Pedraza-Chaverri

Cruz

Tapia

. Activity of serum enzymes in puromycin aminonucleoside-induced nephritic syndrome. Ren Fail 1992; 14: 523–531.

32.

Vlaicu

Popescu

. Serum electrophorenic lipoprotein fators and pseudocholinesterase activity in thyroid disease. Endocrinologie 1978; 16: 147–151.

33.

Osada

López-Miranda

Sastre

. The human hepatoma cell line, HepG2, secretes functional cholinesterase. Biochem Mol Biol Int 1994; 33: 1099–1105.

34.

Scriver

Beaudet

Sly

. Pharmacogenetics in the metabolic & molecular basis of inherited disease, 8th edn. New York, NY: McGraw-Hill, 2001, pp. 234–234.

35.

Marchesini

Brizi

Bianchi

. Nonalcoholic fatty liver disease: a feature of the metabolic syndrome. Diabetes 2001; 50: 1844–1850.

36.

Balkau

Lange

Vol

. Nine-year incident diabetes is predicted by fatty liver indices: the French D.E.S.I.R. study. BMC Gastroenterol 2010; 10: 56–56.

37.

Ratziu

Giral

Charlotte

. Liver fibrosis in overweight patients. Gastroenterology 2000; 118: 1117–1123.

38.

Kotronen

Peltonen

Hakkarainen

. Prediction of non-alcoholic fatty liver disease and liver fat using metabolic and genetic factors. Gastroenterology 2009; 137: 865–872.

39.

Shibata

Kihara

Taguchi

. Nonalcoholic fatty liver disease is a risk factor for type 2 diabetes in middle-aged Japanese men. Diabetes Care 2007; 30: 2940–2944.

40.

Schwenzer

Springer

Schraml

. Non-invasive assessment and quantification of liver steatosis by ultrasound, computed tomography and magnetic resonance. J Hepatol 2009; 51: 433–445.

41.

Fabbrini

Conte

Magkos

. Methods for assessing intrahepatic fat content and steatosis. Corr Opin Clin Nutr Metab Care 2009; 12: 474–481.

42.

Joseph

Saverymuttu

al-Sam

. Comparison of liver histology with ultrasonography in assessing diffuse parenchymal liver disease. Clin Radiol 1991; 43: 26–31.

43.

Jee

Foong

Hur

. Smoking and risk for diabetes incidence and mortality in Korean men and women. Diabetes Care 2010; 33: 2567–2572.

44.

Fehér

Lengyel

. A new approach to drug therapy in non-alcoholic steatohepatitis (NASH). J Int Med Res 2003; 31: 537–551.