Association of liver enzymes with metabolic syndrome and carotid atherosclerosis in young adults. The Cardiovascular Risk in Young Finns Study

Abstract

Objective. We examined whether metabolic syndrome (MetS) predicts increased alanine aminotransferase (ALT) and gamma-glutamyltransferase (GGT) levels in young adults, whether spontaneous recovery from MetS has a favorable effect on liver enzyme activities, and whether these enzymes contribute to the atherogenicity of MetS (assessed by carotid intima-media thickness (IMT)).

Methods. The study included 1,553 subjects (base-line age 31.5 ± 5.0 years). ALT and GGT were measured in 2007. MetS was diagnosed by the new Joint Interim Societies definition.

Results. ALT and GGT levels were higher in subjects with MetS compared to those without in 2007. The association was independent of alcohol intake and BMI. In multivariable models adjusted for base-line age, LDL cholesterol, CRP, alcohol intake, and adiponectin, MetS in 2001 predicted increased ALT (β ± SEM = 0.320 ± 0.062, P < 0.0001 in men; 0.134 ± 0.059, P = 0.02 in women) and GGT (β ± SEM = 0.222 ± 0.067, P < 0.0001 in men; 0.236 ± 0.060, P < 0.0001 in women) levels after 6 years. Subjects with MetS only at base-line (2001) had lower ALT levels after 6 years compared to subjects with persistent and incident MetS. No statistically significant interaction for MetS*ALT (P = 0.81) or MetS*GGT (P = 0.92) on IMT was observed.

Conclusion. In young adults MetS may induce liver enzyme changes that indicate increased risk of non-alcoholic fatty liver disease, but we found no evidence that increased enzyme levels would amplify the atherogenicity of MetS.

Key Words::

• Metabolic syndrome
• non-alcoholic fatty liver disease
• subclinical atherosclerosis

Abbreviations
ALT	=	alanine aminotransferase
BMI	=	body mass index
CRP	=	C-reactive protein
CV	=	coefficient of variation
CVD	=	cardiovascular disease
GGT	=	gamma-glutamyltransferase
HDL	=	high-density lipoprotein
IMT	=	carotid intima-media thickness
LDL	=	low-density lipoprotein
MetS	=	metabolic syndrome
NAFLD	=	non-alcoholic fatty liver disease

Key messages

• Metabolic syndrome (MetS) predicts increased risk of liver fat accumulation assessed by liver enzymes.

• Recovery from the MetS may induce beneficial changes in the liver.

• Increased liver enzyme levels within normal range do not amplify the atherogenicity of MetS.

Introduction

Metabolic syndrome (MetS) is a constellation of several interrelated cardio-metabolic risk factors including obesity, hypertension, dyslipidemia, hyperglycemia, and insulin resistance often accompanied with hyperinsulinemia (1). MetS has been associated with an increased risk of type 2 diabetes, cardiovascular disease (CVD), and fatty liver (2,3). It has been suggested that fatty liver may be the hepatic manifestation of MetS (4).

Non-alcoholic fatty liver disease (NAFLD) is one of the most common causes of liver disease in adults (5). Fatty liver in NAFLD may proceed to steatohepatitis and further to cirrhosis and liver failure (6). Elevated alanine aminotransferase (ALT) and gamma-glutamyltransferase (GGT) are markers of liver damage that have been associated with fat accumulation in liver and CVD (4,7). Recently, fatty liver assessed by elevated liver enzymes has been found to relate with carotid intima-media thickness (IMT)—a marker of subclinical atherosclerosis (8)—independently of MetS (9). In addition it has been suggested that the association between NAFLD and IMT reflects adverse effects of MetS (10).

However, data are lacking on whether MetS is associated with liver ALT or GGT levels in young adults and whether these enzymes indicate hepatic contribution to the atherogenicity of MetS. Therefore, the objectives of this study were to 1) determine the cross-sectional association of MetS with ALT and GGT; 2) to explore whether MetS at base-line (at ages 24–39 years in 2001) is related with liver enzymes at follow-up; 3) to determine whether recovery from the MetS after 6 years is associated with liver enzyme activities; and 4) to study the associations of liver enzymes with IMT. The study was conducted among 1,553 men and women from the population-based Cardiovascular Risk in Young Finns Study for whom ALT and GGT levels were available in 2007.

Materials and methods

Subjects

The Cardiovascular Risk in Young Finns Study is an on-going epidemiological study to assess biological and life-style factors underlying CVD. The first cross-sectional survey was conducted in 1980 when 3,596 randomly selected children and adolescents (age 3–18 years) participated. Thereafter, several follow-up studies have been performed, with the most recent in 2001 and 2007. The sample for the present analysis included those subjects who had ALT and GGT levels measured from the 2007 follow-up study and had full metabolic risk data from both 2001 and 2007 follow-ups. Liver enzyme data from follow-up in 2001 was not available. In 2007, 198 (9.3%) subjects that were considered as heavy drinkers (alcohol intake > 20 g/day for women and > 30 g/day for men) (11) were excluded from the analysis. A total of 629 men and 924 women were included in the present analysis. BMI data were available for 1,539 subjects and alcohol intake data for 1,532 subjects. Subjects gave written informed consent, and the study was approved by local ethics committees.

Clinical characteristics

Anthropometric measurements, blood pressure, and questionnaires. Height, weight, and waist circumferences were measured. Waist circumference was assessed midway between the iliac crest and the lowest rib as the average of two measurements with an accuracy of 0.1 cm. Body mass index (BMI) was calculated using the formula: weight (kg)/(height (m))². Blood pressure was measured by using a random zero sphygmomanometer. The average of three measurements was used in the analysis. Smoking habits and alcohol intake were inquired with the use of questionnaires (12). Subjects were asked to report their alcohol consumption of cans or bottles (1/3 L) of beer, glasses (12 cL) of wine, and shots (4 cL) of strong alcohol per week. The values of different beverages consumed during the week allowed us to determine the total alcohol intake per day.

ALT and GGT measurements. Venous blood samples were collected after a 12-hour fast. ALT and GGT activities were measured enzymatically (ALT and GGT System Reagent, Olympus, Ireland) on an automatic analyzer (AU400, Olympus, Japan). The mean inter-assay coefficient of variation (CV) was 3.7% and 2.1%, respectively. We defined aminotransferase elevation i.e. NAFLD (men n = 70 and women n = 38) in the present study as any value above normal of ALT based on the definition suggested by Prati et al. (13); for men this corresponded to ALT > 40 U/L, and for women ALT > 30 U/L.

Biomarkers of atherosclerosis. Determination of serum glucose, triglycerides, total cholesterol, and high-density lipoprotein (HDL) cholesterol was done by automated, enzymatic methods as described previously (12). Low-density lipoprotein (LDL) cholesterol was calculated using the Friedewald formula (14) for subjects with triglycerides < 4 mmol/L. Serum insulin concentrations were determined with an IMx analyzer (CV 2.1%) (Abbott Laboratories, USA) by microparticle enzyme immunoassay. C-reactive protein (CRP) was determined turbidimetrically (2001: CRP-UL reagent, Wako, USA; 2007: CRP Latex reagent, Olympus) on an AU400 analyzer (Olympus). The inter-assay CV was 3.3%. Radioimmunoassay was used to measure concentrations of serum adiponectin with an inter-assay CV of 11.9% (Human Adiponectin RIA kit; Linco Research, St Charles, Missouri, USA).

Definition of MetS

Subjects were considered having MetS when the diagnostic criterion of the Joint Interim Statement of the International Diabetes Federation Task Force on Epidemiology and Prevention, National Heart, Lung, and Blood Institute, American Heart Association, World Heart Federation, International Atherosclerosis Society, and International Association for the Study of Obesity definition of MetS was fulfilled (15). MetS was identified when three or more of the following five criteria were met: waist circumference ≥ 102 cm in men or ≥ 88 cm in women; triglycerides ≥ 1,695 mmol/L; HDL cholesterol < 1.036 mmol/L in men or < 1.295 mmol/L in women; blood pressure ≥130/≥85 mmHg or on antihypertensive medication; fasting glucose ≥ 5.6 mmol/L.

Subjects were classified further into four groups according to their MetS status at both time points (2001 and 2007): recovery group (MetS at base-line but not at follow-up, n = 72), incident group (MetS at follow-up but not at base-line, n = 118), persistent group (MetS both at base-line and at follow-up, n = 112), and control group (no MetS at base-line or at follow-up, n = 1,251) (16).

Ultrasound imaging

Ultrasound studies were performed by trained physicians and sonographers following standardized protocols in five centers. Mean IMT was derived using a minimum of four IMT measurements from the posterior (far) wall of the left carotid artery approximately 10 mm proximal to the carotid bifurcation. Measurements were made afterwards from stored digital images by one experienced reader (M.J.) blinded to the subjects’ clinical characteristics. An ultrasound imaging device with a high-resolution system (Sequoia 512; Acuson, CA, USA) was used. The coefficient of variation between visits was 6.4%. Full details of the methods have been described previously (16,17).

Statistical methods

Values for ALT, GGT, triglycerides, insulin, CRP, and adiponectin were log_e-transformed to correct for skewness. Cross-sectional analysis was based on data in 2007. Linear regression was used to examine for significant study group*sex interactions for liver enzyme measures. If no significant interactions were observed, we did not stratify the analysis by sex. Otherwise, the analyses were performed stratified by sex.

Characteristics of the study subjects were compared using the t test and chi-square test as appropriate. Analysis of covariance was used to assess liver enzyme levels between subjects with and without MetS as well as to calculate mean ALT and GGT activities across groups according to the number of MetS components (0, 1, 2, 3, and > 4 components). Pearson's correlation coefficients adjusted for age were calculated to assess bivariate associations between liver enzymes and risk factors. To study the independent relations between MetS and liver enzymes, we first calculated standardized values (z score) at base-line and follow-up for risk components. Second, we performed multivariable regression adjusting for components not included in the MetS definition (age, BMI, LDL cholesterol, CRP, alcohol intake, and adiponectin). Due to high multicollinearity, MetS and its components were not included in the same multivariable analyses. Analysis of covariance (age and sex as covariates) was used to compare mean liver enzyme levels of recovery to incident, persistent, and control groups. Linear regression was used to examine for significant ALT and GGT*MetS interactions on IMT. Age-, sex-, alcohol intake-, and MetS-adjusted analysis of covariance was used to examine the mean IMT levels in subjects with and without NAFLD. Multivariable regression was used to examine the associations between IMT and liver enzymes at follow-up. Statistical analyses were performed with SAS version 9.1, and statistical significance inferred as a two-tailed P value ≤ 0.05.

Results

Clinical characteristics

In 2007, a total of 149 subjects (9.5%) were taking antihypertensive medications, and 45 subjects (2.9%) were taking lipid-lowering medication. In the present study 15 subjects reported having insulin-treated type 1 diabetes. The results presented were essentially similar when subjects with these medications and type 1 diabetes were excluded from the analyses. Characteristics for study subjects at base-line in 2001 are summarized in Table I. The levels of body composition, blood pressure, lipids (except for HDL cholesterol), IMT, and other biochemical parameters were higher in subjects with MetS compared to those without. No difference between these groups were observed in alcohol intake or smoking.

Table I. Characteristics of the study population at base-line in 2001.

	Men		Women
Variable 2001	MetS(+)	MetS(−)	MetS(+)	MetS(−)
Number of subjects	100	529	84	840
Age (years)	33.3 ± 5.0	31.6 ± 5.0	33.0 ± 4.8	31.8 ± 5.0
Waist circumference (cm)	102 ± 11	87 ± 9	98 ± 11	77 + 9
Body mass index (kg/m^2)	29.6 ± 4.4	24.8 ± 3.3	31.9 ± 5.1	23.6 ± 3.7
Systolic blood pressure (mmHg)	130 ± 12	119 ± 11	124 ± 17	111 ± 11
Diastolic blood pressure (mmHg)	82 ± 11	71 ± 10	78 ± 12	68 ± 9
Total cholesterol (mmol/L)	5.63 ± 0.96	5.14 ± 0.97	5.50 ± 0.97	4.99 ± 0.85
LDL cholesterol (mmol/L)	3.59 ± 0.91	3.40 ± 0.89	3.51 ± 0.83	3.10 ± 0.74
HDL cholesterol (mmol/L)	0.94 ± 0.19	1.19 ± 0.27	1.10 ± 0.27	1.41 ± 0.28
Triglycerides (mmol/L)	2.33 (1.77–2.99)	1.15 (0.80–1.50)	1.80 (1.30–2.50)	0.97 (0.80–1.20)
Glucose (mmol/L)	5.65 ± 1.61	5.08 ± 0.50	5.35 ± 0.77	4.89 ± 0.78
Insulin (mU/L)	11.7 (8.0–16.0)	5.7 (4.0–8.0)	13.2 (9.0–18.0)	6.1 (4.0–8.0)
CRP (mg/L)	1.30 (0.63–2.45)	0.57 (0.26–1.12)	2.51 (1.60–4.21)	0.80 (0.34–1.91)
Adiponectin (ug/mL)	5.55 (3.80–6.70)	7.60 (5.4–9.3)	8.42 (5.9–9.7)	11.19 (8.10–13.85)
Daily smoking (%)	23	24	12	19
Alcohol intake (drinks/day)	1.2 (0.2–1.7)	1.2 (0.3–1.7)	0.5 (0–0.7)	0.5 (0–0.7)
Carotid IMT (mm)	0.62 ± 0.12	0.58 ± 0.10	0.60 ± 0.08	0.57 ± 0.09

Comparisons between subjects with MetS (+) and those without MetS (−).

P always < 0.001, except for alcohol intake (P = 0.90 for men and P = 0.28 for women) and smoking (P = 0.79 for men and P = 0.14 for women).

Values are mean ± SD or geometric mean (25–75 percentiles) for continuous variables and percent for smoking.

Cross-sectional analysis

A total of 230 subjects (14.8%) (n = 131 men and n = 99 women) had MetS at the 2007 follow-up. Age-and sex-adjusted ALT and GGT levels in 2007 were higher in subjects with MetS compared to those without MetS (Table II). To assess whether the association of MetS on liver enzyme levels were mediated by alcohol intake or obesity, we compared the levels of ALT and GGT according to MetS status in subgroups of similar alcohol intake or BMI in 2007. The age-and sex-adjusted results are displayed in Table II. Subjects with MetS had higher liver enzyme activities compared to those without MetS at all levels of alcohol intake and BMI status. To further assess whether MetS is cross-sectionally associated with increased liver enzymes, regression models that included standardized age, LDL cholesterol, CRP, alcohol intake, and adiponectin were constructed. MetS was associated in both men and women with ALT (β = 0.31 ± 0.06, P < 0.0001; and b = 0.29 ± 0.06, P < 0.0001, respectively) and GGT (β = 0.26 ± 0.06, P < 0.0001; and β = 0.26 ± 0.06, P < 0.0001, respectively). The results were similar when BMI or insulin was included into the models.

Table II. Cross-sectional age-and sex-adjusted comparison of liver enzyme levels between subjects with MetS (+) and without MetS (−) in populations with similar alcohol intake and BMI in 2007. Data from follow-up in 2007.

		ALT			GGT
		MetS(−)	MetS(+)	P value	MetS(−)	MetS(+)	P value
Mean (U/L)		13.9 (10.0–19.0)	23.0 (14.0–25.0)	<0.0001	20.8 (15.0–35.0)	36.9 (22.0–61.0)	<0.0001
Drinks per day < 1	n = 1099	13.0 (10.0–19.0)	18.9 (13.0–26.0)	<0.0001	18.8 (13.0–25.0)	27.9 (19.0–48.0)	<0.0001
Drinks per day 1 to < 2	n = 346	14.9 (12.0–22.0)	24.3 (23.0–52.0)	<0.0001	22.6 (15.0–30.0)	38.9 (30.0–66.0)	<0.0001
Drinks per day > 2	n = 87	14.3 (14.0–34.0)	23.6 (23.0–52.0)	0.004	27.1 (17.0–38.0)	38.9 (22.0–76.0)	0.06
BMI < 25 kg/m^2	n = 732	11.8 (9.0–16.0)	15.2 (13.0–23.0)	0.04	16.8 (12.0–23.0)	27.8 (14.0–50.0)	<0.0001
BMI 25 to < 30 kg/m^2	n = 540	15.2 (10.0–26.5)	18.3 (13.0–26.5)	0.002	23.2 (14.0–35.0)	29.6 (21.0–66.0)	0.002
BMI > 30 kg/m^2	n = 267	18.7 (12.5–25.5)	23.6 (16.0–39.0)	0.003	29.1 (18.0–40.0)	34.4 (23.0–58.0)	0.03

Prospective analysis

Age-adjusted correlation coefficients for liver enzymes in 2007 with base-line 2001 MetS, risk factors, and alcohol intake are shown in Table III. The strongest correlates of future liver enzymes were adiposity measures and MetS in both sexes.

Table III. Age-adjusted Pearson correlation coefficients for bivariate associations between risk variables measured in 2001 and liver enzymes 6 years later.

	Men		Women
Variable 2001	Log-ALT	Log-GGT	Log-ALT	Log-GGT
MetS	0.26	0.21	0.13	0.20
Waist	0.37	0.36	0.28	0.33
BMI	0.34	0.32	0.29	0.33
Systolic blood pressure	0.17	0.16	0.11	0.11
Diastolic blood pressure	0.15	0.19	0.08^c	0.13
Total cholesterol	0.15	0.22	0.08^c	0.12
LDL cholesterol	0.11	0.17	0.09^c	0.09^c
HDL cholesterol	−0.19	−0.17	−0.06^b	−0.04^a
Log-Triglycerides	0.28	0.33	0.09	0.18
Glucose	0.11	0.07^b	0.05^b	0.07^b
Log-Insulin	0.30	0.25	0.19	0.24
Log-CRP	0.20	0.26	0.12	0.19
Alcohol intake	−0.01^a	0.06^a	0.05^a	0.13

All P values < 0.001 unless informed otherwise.

^aP > 0.12.

^bP < 0.05.

^cP < 0.01.

Results for multivariable regression analyses in men and women evaluating the independent base-line determinants in predicting increased liver enzymes activities 6 years later are shown in Table IV. Base-line MetS was associated in both men and women with ALT and GGT at follow-up in 2007. The results were similar when further adjusted for concurrent alcohol intake (2007). When BMI was included into the models as a dependent variable, the association in women between MetS and ALT (P = 0.42) and between MetS and GGT (P = 0.30) 6 years later was diminished. When BMI was replaced with base-line insulin, the association between MetS and GGT in women was diminished (P = 0.27). Next we assessed whether ALT and GGT levels in 2007 were associated with increasing number of MetS components diagnosed in 2001 (central obesity, hypertension, triglyceridemia, low HDL cholesterol, hyperglycemia). To examine this, we plotted liver enzymes against the number of MetS components. There was an increasing trend in ALT (P for trend < 0.0001) and GGT (P for trend < 0.0001) activity with increasing number of base-line MetS components (age, sex, BMI, LDL cholesterol, CRP, alcohol intake, and adiponectin adjusted).

Table IV. Multivariable associations between risk variables in 2001 and liver enzymes in 2007. Values are standardized regression coefficients (expressed in U/I) for 1 SD change in explanatory variables.

	Log-ALT (2007)		Log-GGT (2007)
	Beta ± SE	P value	Beta ± SE	P value
Men:
MetS	0.320 ± 0.062	<0.0001	0.222 ± 0.067	<0.0001
Age	−0.003 ± 0.022	0.89	0.055 ± 0.023	0.02
LDL cholesterol	0.051 ± 0.020	0.01	0.088 ± 0.022	0.0009
Log-CRP	0.077 ± 0.023	0.01	0.128 ± 0.025	<0.0001
Alcohol intake	0.005 ± 0.017	0.78	0.041 ± 0.018	0.03
Log-Adiponectin	−0.080 ± 0.026	0.002	−0.110 ± 0.028	<0.0001
Women:
MetS	0.134 ± 0.059	0.02	0.236 ± 0.060	<0.0001
Age	0.012 ± 0.017	0.45	0.063 ± 0.017	<0.0001
LDL cholesterol	0.042 ± 0.018	0.02	0.041 ± 0.018	0.02
Log-CRP	0.040 ± 0.016	0.02	0.066 ± 0.016	<0.0001
Alcohol intake	0.049 ± 0.021	0.02	0.110 ± 0.022	<0.0001
Log-Adiponectin	−0.061 ± 0.018	0.0007	−0.073 ± 0.018	<0.0001

To assess liver enzyme activities after recovery from the MetS, we compared age and sex-adjusted levels of liver enzymes between subjects with base-line (in 2001) only MetS and those with incident (only at follow-up in 2007) or with persistent (both at base-line and follow-up) MetS. In 2007, subjects in the recovery group had significantly lower ALT activities compared to incident and persistent group (Figure 1). No difference between recovery and incident groups (P = 0.17) or between recovery and persistent groups (P = 0.17) were observed in GGT. Subjects in the control group had significantly lower GGT activities compared to recovery group (mean, 95% confidence intervals 19.6, 13.0–26.0 versus 27.4, 18.0–38.5 U/L; P < 0.0001).

Figure 1. Age-and sex-adjusted mean and 95% confidence intervals for ALT activity in recovery group versus incident, persistent, and control group.

The relations of liver enzymes on IMT in subjects with and without MetS

Multivariable regression model was used to examine cross-sectional association (in 2007) between liver enzymes and IMT. ALT and GGT were associated with IMT independent of age, sex, and alcohol intake (Table V). When MetS was introduced into the models, the association was diluted to non-significant as shown in Table V. Next, to examine association between MetS and NAFLD in 2007, we compared IMT in subjects with and without NAFLD. The model was adjusted for age, sex, alcohol intake, and MetS. No difference in IMT between subjects with NAFLD and subjects without it was observed (0.623 ± 0.002 versus 0.636 ± 0.009; P = 0.35). Results were similar when MetS was replaced by BMI (data not shown). Further, to examine the associations between liver enzymes and IMT according to the presence of MetS, we first examined whether cross-sectional interaction between liver enzymes and MetS on IMT was present. No statistically significant interaction for MetS*ALT (P = 0.81) or MetS*GGT (P = 0.92) on IMT was observed. Finally, IMT was regressed against liver enzyme activity stratified by MetS status. Models were adjusted for age, sex, alcohol intake, and BMI. No linear association between ALT and IMT (β = 0.004 ± 0.012, P = 0.71 in subjects with MetS; and β = 0.0001 ± 0.04, P = 0.99 in subjects without MetS) or between GGT and IMT (β = 0.006 ± 0.010, P = 0.57 in subjects with MetS; and β = 20.005 ± 0.004, P = 0.29 in subjects without MetS) was observed. Results were similar when models were stratified by MetS status in 2001 (data not shown).

Table V. Multivariable cross-sectional association between liver enzymes and IMT in 2007.

	Model 1^a		Model 2^b
	β ± SE	P value	β ± SE	P value
ALT	0.015 ± 0.004	0.0005	0.008 ± 0.004	0.06
GGT	0.013 ± 0.004	0.002	0.006 ± 0.004	0.16

^aAdjusted for age, sex, and alcohol use.

^bFurther adjustment for MetS.

Discussion

The present data demonstrate that MetS was independently associated with increased liver enzyme activity both cross-sectionally and prospectively when adjusting for components not included in MetS definition. Recovery from MetS over a 6-year period was associated with lower activities of ALT at the follow-up compared to those with incident or persistent MetS. Liver enzyme activity did not modify the association between MetS and IMT.

Subjects with MetS had increased levels of ALT and GGT (compared to those without MetS) independent of obesity and alcohol intake suggestive of liver changes that may indicate increased liver fat content (4,18). Patel et al. showed in a large group of young adults that elevation of ALT and GGT, as biomarkers of liver dysfunction and NAFLD, associated adversely to MetS and its components as well as to history of coronary artery disease (19). These data are in line with previous studies in which the association of MetS with biopsy, ultrasound, and magnetic resonance spectroscopy-proven NAFLD was observed (4). Kotronen et al. (4) reported that in subjects (aged 20–65 years) with MetS, liver fat content was 4-fold higher (8.2% versus 2.0%) compared to subjects without MetS. Although fatty liver has been associated strongly to alcohol abuse and obesity, we showed in the present study that when subjects within a narrow range of BMI and with similar alcohol drinking habits were studied, the association between MetS and liver enzyme levels remained significant. This is in agreement with studies showing that not all morbidly obese individuals develop NAFLD (20) and that some individuals with a fatty liver are not obese (21). Thus, obesity does not seem to have a direct role in promoting liver fat but may intensify the effects of other factors predisposing liver fat accumulation (18). Further, we observed an increasing trend in serum liver enzyme activities with increasing number of metabolic components. A similar trend has been observed in a small-scale clinical study that examined the association between MetS components and liver fat assessed by magnetic resonance spectroscopy (22). These observations may suggest that, in the presence of MetS, elevated liver enzyme activity even within normal range may indicate the severity of the syndrome, thus accelerating the development of clinical NAFLD.

In the present study, MetS at base-line correlated independently with ALT and GGT at follow-up. However, we observed that when adjusting for BMI or insulin the association between MetS and ALT 6 years later was diminished in women. This observation may be explained by female reproductive hormones that could contribute to lower serum ALT concentrations (23) but may also indicate that increased ALT activity in women is associated with insulin resistance. Insulin resistance is considered as a key underlying factor in the development of MetS (1). Suppression of the inhibitory effect of insulin on lipolysis in adipose tissue may lead to an increased production of free fatty acids and glucose production from the liver, thus inducing accumulation of fat in liver (24).

To our knowledge, this is the first study to show that a spontaneous recovery from MetS over a 6-year period is associated with lower serum ALT levels compared to those who had incident or persistent MetS at follow-up. In line with this observation, several intervention studies have reported reduced levels of liver fat deposits after thiazolidinedione treatment and favorable developments in life-style habits (4). However, we did not have data on liver enzymes available from the base-line (in 2001) examination and therefore were not able to evaluate whether there were base-line differences between study groups.

Increased IMT is associated with conventional CVD risk factors, relates to severity of CVD, and predicts CVD events (8). We (25) and others (26) have previously shown that young adults with MetS have accelerated IMT progression. MetS and NAFLD frequently coexist (27). In cross-sectional settings NAFLD has been found to correlate with IMT even independent of MetS (9). Pacifico et al. showed that obese children with NAFLD had significantly increased carotid IMT than obese children without liver involvement and controls. The severity of fatty liver as assessed by ultrasound was independently associated with IMT (28). In addition, two large longitudinal studies that included middle-aged adults have demonstrated a dose–response association between liver enzymes and cardiovascular events (29,30). Fatty liver and atherosclerosis may share common molecular mediators (31). However, it is somewhat unclear whether NAFLD is merely a marker or an early mediator in the development of atherosclerosis (31). In the present study, the association between liver enzymes and IMT was diluted when taking MetS into account, suggesting that liver changes assessed by ALT and GGT may not indicate an increased risk of CVD over and above what would be expected due to the increased risk of MetS. Our results are in line with previous literature regarding CVD risk in subjects with rigorously determined NAFLD (32). However, our findings were limited to cross-sectional design, and therefore causal association between elevated liver enzymes and IMT could not be studied. Further, Kim et al. recently reported that after adjusting for conventional cardiovascular risk factors, the independent association between NAFLD and IMT was limited to subjects with MetS or multiple metabolic abnormalities (10). However, we observed no MetS*ALT or MetS*GGT interaction on IMT. These data suggest that increased liver enzyme activities in young adults may not be indicative of additive risk of atherosclerosis among subjects with MetS.

Our study had limitations. Currently, the available chemical markers of fatty liver in clinical practice are somewhat limited (4). In the present study, we were not able to measure liver fat content with ascertainable methods such as magnetic resonance spectroscopy or biopsy. Liver enzyme levels fluctuate over time; therefore one time measurement may under-estimate liver fat content (33). We did not have data on liver enzymes available from the base-line (in 2001) examination and therefore were not able to evaluate whether there were base-line differences between study groups. In addition, we were not able to exclude subjects with viral (hepatitis B and C), toxic, autoimmune, or other rare causes of elevated liver enzymes (Wilson disease, hemochromatosis) from the cohort, which may have had an influence on results. Because the mean values of increased liver enzymes were in the upper limit of normal range, the results should be interpreted with caution. Because our study cohort was racially homogeneous, the generalizability of our results is limited to white European subjects. The strength of this study was the large randomly selected cohort of ambulatory young adults and follow-up data of MetS.

In conclusion, the present data demonstrate in young adults that MetS is associated cross-sectionally and prospectively with elevated liver enzyme activities even within normal range that may be indicative of increased risk of NAFLD later in life. Increased liver enzyme levels may be restored after recovery from MetS. However, while recovery from MetS was associated with beneficial changes in liver enzyme levels, we observed that subjects in the recovery group continued to exhibit elevated ALT and GGT compared to ambulatory controls without MetS at base-line and follow-up. This finding suggests that individuals with a past history of MetS are in need for continued vigilance for adverse health outcomes even after apparent recovery from MetS. Finally, we found no evidence that increased liver enzyme levels within normal range would indicate amplified risk of atherosclerosis among subjects with MetS.

Declaration of interest: This study was financially supported by the Academy of Finland (grant no. 117797, 121584, 126925), Turku University, the Turku University Foundation, the Tampere University Hospital Medical Fund, the Finnish Foundation of Cardiovascular Research, the Lydia Maria Julin Foundation, Juho Vainio Foundation, Finnish Cultural Foundation, Finnish Medical Foundation, and Research Foundation of Orion Corporation. The authors declare no conflicts of interest.