Liver Fat Is Increased in Type 2 Diabetic Patients and Underestimated by Serum Alanine Aminotransferase Compared With Equally Obese Nondiabetic Subjects Free

A total of 70 nondiabetic subjects and 70 type 2 diabetic patients were recruited by newspaper advertisements and by contacting occupational health services in Helsinki based on the following inclusion criteria: 1) aged 18–70 years; 2) no known acute or chronic disease based on history, physical examination, and standard laboratory tests (blood counts, serum creatinine, thyroid-stimulating hormone, electrolyte concentrations, and electrocardiogram); 3) alcohol consumption <20 g/day; and 4) no evidence of hepatitis A, B, or C or of autoimmune hepatitis or clinical signs or symptoms of inborn errors of metabolism or history of use of toxins or drugs known to induce hepatitis. Exclusion criteria included proliferative retinopathy and use of antihypertensive agents that could possibly influence glucose metabolism (β-blockers and thiazides) or of thiazolidinediones. A total of 13 nondiabetic subjects and 32 type 2 diabetic patients received antihypertensive medications (ACE inhibitors or calcium channel blockers). A total of 31 type 2 diabetic patients were treated with diet alone, 17 were treated with metformin, and 22 were treated with a combination of metformin and insulin. Five nondiabetic subjects and 22 type 2 diabetic patients were receiving statins. For the insulin-treated patients, additional inclusion criteria included stable body weight and insulin dose for at least 6 months. Data for nondiabetic subjects (9) and for 46 of the type 2 diabetic patients participating in treatment studies (5,6,11) have been reported previously. For the latter patients, only baseline data are included. The study protocol was approved by the ethics committee of the Helsinki University Central Hospital, and each participant provided written informed consent.

Liver fat content was measured by proton magnetic resonance spectroscopy as described previously (5). This measurement has been validated against histologically determined lipid content (12) and against estimates of fatty degeneration or infiltration by X-ray computer-assisted tomography (5). All spectra were analyzed by a physicist in a blinded fashion. The reproducibility of repeated measurements of liver fat in nondiabetic subjects studied on two occasions in our laboratory is 11% (13).

Intra-abdominal and subcutaneous fat content were determined by magnetic resonance imaging as described previously (5). Percent body fat was determined using bioelectrical impedance spectroscopy (14). Waist circumference was measured midway between the spina iliaca superior and the lower rib margin (15).

Plasma glucose, A1C, serum free insulin, fasting serum C-peptide, fasting serum HDL cholesterol, fasting serum triglyceride, fasting serum LDL, fasting serum free fatty acids, serum aspartate aminotransferase (S-AST), and S-ALT concentrations were measured as described previously (9).

Non-normally distributed data were used after logarithmic (base 10) transformation. If distributed normally, data are shown as means ± SEM, whereas non-normally distributed data are shown as median and the interquartile range (25th –75th percentiles). An unpaired Student's t test was used to compare mean values between the groups. ANCOVA was used to compare slopes and intercepts of regression lines for the nondiabetic subjects and the type 2 diabetic patients and for comparison of liver fat in the type 2 diabetic patients treated with diet, with metformin, or with insulin-metformin combination therapy. If neither the slopes nor the intercepts differed between the groups, a common regression equation was calculated. Correlation analyses were performed using Spearman's nonparametric rank correlation coefficient. ANCOVA was used to adjust for intra-abdominal fat content. Calculations were performed using GraphPad Prism (version 4.00 for Windows; GraphPad Software, San Diego, CA), and SPSS 14.0 for Windows (SPSS, Chicago, IL). P < 0.05 was considered statistically significant.

Nondiabetic subjects and type 2 diabetic patients were equally obese and similar with respect to age and sex (Table 1). The type 2 diabetic patients had higher fasting serum insulin, fasting serum triglyceride, and fasting serum free fatty acid concentrations and lower fasting serum HDL cholesterol concentrations that the nondiabetic subjects. The type 2 diabetic patients had ∼16% more intra-abdominal and ∼80% more liver fat than the nondiabetic subjects. The difference in liver fat remained significant even after adjustment for intra-abdominal fat content (P < 0.05). The type 2 diabetic patients treated with diet, metformin, or the combination of insulin and metformin differed slightly with respect to age (45 ± 2 [mean ± SEM], 55 ± 2, and 50 ± 2 years for diet, metformin, and insulin and metformin, respectively; P = 0.003 for diet versus metformin), whereas BMI (32.0 ± 0.7, 30.2 ± 1.4, and 32.2 ± 1.0 kg/m²; NS), liver fat content (median 15.5% [interquartile range 7.0–21.0], 9.0% [4.2–19.5], and 14.0% [7.0–21.3]; NS), and S-ALT concentrations (42 [29–69], 31 [21–46], and 38 units/l [26–52] NS) were similar between the groups.

Liver fat content was significantly correlated with BMI and waist circumference in the type 2 diabetic patients and in the nondiabetic subjects (Fig. 1). At any given BMI or waist circumference, the type 2 diabetic patients had significantly more liver fat than the nondiabetic subjects. The type of antihyperglycemic therapy had no impact on these relationships (data not shown). At a BMI of 25 kg/m², the type 2 diabetic patients had ∼40% more liver fat than the nondiabetic subjects. This difference rose gradually as a function of BMI. For example, at a BMI of 40 kg/m², the type 2 diabetic patients had 80% more liver fat than the nondiabetic subjects. A similar phenomenon was observed when waist was plotted against liver fat. For example, at waist circumferences of 85 and 140 cm, liver fat content was 50 and 90% higher in the type 2 diabetic patients than in the nondiabetic subjects, respectively. Intra-abdominal fat was similarly related to liver fat in both nondiabetic subjects and type 2 diabetic patients (r = 0.45, P < 0.0001). The volume of subcutaneous fat did not correlate with liver fat content in either group.

Liver fat content was similarly related to fasting serum insulin (r = 0.55, P < 0.0001), C-peptide (r = 0.40, P < 0.0001), A1C (r = 0.34, P < 0.0001), fasting plasma glucose (r = 0.29, P = 0.0006), fasting serum triglyceride (r = 0.36, P < 0.0001), and fasting serum HDL cholesterol (r = −0.31, P = 0.0002) concentrations in the nondiabetic subjects and the type 2 diabetic patients (Fig. 2).

S-AST and S-ALT correlated with liver fat content in both nondiabetic subjects and type 2 diabetic patients. The slopes of the regression lines between S-ALT concentrations and liver fat content differed significantly between the nondiabetic subjects and the type 2 diabetic patients (P = 0.004). For any given S-ALT, the type 2 diabetic patients had more liver fat than the nondiabetic subjects (Fig. 3). Liver fat content did not differ between the groups at low S-ALT concentrations (10–20 units/l) but was 70, 125, and 200% higher in type 2 diabetic patients compared with control subjects at S-ALT concentrations of 50, 100, and 200 units/l. At normal S-AST concentrations (15–45 units/l), liver fat content was 30–70% higher in the type 2 diabetic patients than in the nondiabetic subjects. This increase rose gradually with increasing S-AST concentrations and averaged 80 and 130% at S-AST concentrations of 60 and 100 units/l. Serum liver enzyme concentrations did not differ between the subjects who were using statins compared with those who were not (S-AST median 29 units/l [interquartile range 22–40] and 29 units/l [23–46], NS; S-ALT 35 [21–51] and 33 units/l [24–64], NS). The relationship between liver enzymes and liver fat content did not differ between subjects who were and were not using statins in either group (data not shown).

In the present study, we used state-of-the-art methodology for quantification of liver fat in a large group of type 2 diabetic patients and carefully age- and sex-matched equally obese nondiabetic subjects. Liver fat content was on average 80% higher in the type 2 diabetic patients than in the nondiabetic subjects. This difference was not influenced by the type of antihyperglycemic treatment. Somewhat unexpectedly, S-ALT and S-AST were not related to liver fat similarly in the diabetic and nondiabetic subjects. Both S-ALT and S-AST underestimated liver fat content in type 2 diabetic patients. This underestimation became more pronounced at increasing concentrations of both S-ALT and S-AST.

In many previous studies, liver fat content, measured by proton magnetic resonance spectroscopy, has been found to exceed the upper limit of normal liver fat content (5.56%) (16) in type 2 diabetic patients. However, except for one study that included 10 nondiabetic control subjects (4), in these studies there was either no control group (5,6,11,17–19) or the control subjects were not matched for body weight (20,21). It has thus remained unclear whether the increase in liver fat content has just been a consequence of obesity in type 2 diabetic patients. The present study suggests that type 2 diabetic patients have more liver fat at any given BMI than nondiabetic subjects and that the difference in liver fat content between the groups increases with increasing obesity. In keeping with Kelley et al. (4), intra-abdominal fat was similarly related to liver fat in both type 2 diabetic patients and nondiabetic subjects, and the type 2 diabetic patients had more intra-abdominal fat than the nondiabetic subjects.

The higher liver fat content in the type 2 diabetic patients could contribute to diabetic dyslipidemia. The fatty liver overproduces VLDL particles in both nondiabetic subjects and type 2 diabetic patients, which, in the face of unchanged VLDL clearance, increases serum triglyceride concentrations (21). Hypertriglyceridemia in turn leads to low HDL cholesterol concentrations (22). The similar relationship between triglycerides and HDL cholesterol concentrations and liver fat content in the diabetic patients and nondiabetic subjects (Fig. 2) combined with the kinetic study (21) suggests that excess liver fat indeed contributes to diabetic dyslipidemia.

An intriguing observation in the present study was that the type 2 diabetic patients had 40–200% more liver fat at the same S-ALT and S-AST concentrations than the nondiabetic subjects (Table 2). Both enzymes thus underestimate liver fat in type 2 diabetes, and for any given amount of liver fat, S-ALT and S-AST are lower in type 2 diabetic patients than in normal subjects. Both S-ALT and S-AST increase in response to hepatocyte damage until hepatocytes are lost, and cirrhosis develops (23,24). To what extent the lower enzyme levels reflect a difference in hepatocellular damage cannot be determined from the present study.

It is important to develop tools to diagnose a fatty liver in type 2 diabetic patients because nonalcoholic steatohepatitis is more common in type 2 diabetic patients than in nondiabetic subjects (25,26) and can progress to cirrhosis and liver failure (27). The present data suggest that if assessed using S-ALT or S-AST, hepatic steatosis is underestimated in type 2 diabetic patients compared with equally obese nondiabetic subjects. Thus, there is a need to develop new serum markers of steatosis to complement S-AST and S-ALT, which have been used in the clinic for almost 50 years (28).

This study was supported by research grants from the Academy of Finland, the Sigrid Juselius Foundation, the Medical Society of Finland, and Novo Nordisk Foundation. This work is part of the project Hepatic and Adipose Tissue and Functions in the Metabolic Syndrome (www.hepadip.org), which is supported by the European Commission as an Integrated Project under the 6th Framework Programme (Contract LSHM-CT-2005-018734).

We gratefully acknowledge Katja Tuominen, Mia Urjansson, and Pentti Pölönen for excellent technical assistance and Anna-Maija Häkkinen, Leena Ryysy, Mirja Tiikkainen, Satu Vehkavaara, and Anneli Seppälä-Lindroos for their contributions.