Fatty liver predicts impaired fasting glucose and type 2 diabetes mellitus in Japanese undergoing a health checkup
Dr Tamaki Yamada, Okazaki City Medical Association, Public Health Center, 1-9-1 Tatsumi-nishi, Okazaki, Aichi 444-0875, Japan. Email: firstname.lastname@example.org
Background and Aim: The question of whether fatty liver might predict impaired fasting glucose or type 2 diabetes mellitus in a longitudinal manner was assessed in Japanese subjects undergoing a health checkup.
Methods: A total of 12 375 individuals (6799 men and 5576 women) without hyperglycemia or type 2 diabetes mellitus in 2000 and participating in 2005 were included. Multiple logistic regression analyses were performed for both sexes, adjusted for age, body mass index, elevated blood pressure or hypertension, family history of diabetes mellitus, alcohol drinking and smoking.
Results: Impaired fasting glucose and type 2 diabetes mellitus were newly diagnosed in 7.6% and 1.0% of men and 3.8% and 0.5% of women, respectively, within the 5-year period. The prevalence of newly diagnosed impaired fasting glucose and type 2 diabetes mellitus was significantly higher in the participants with fatty liver than without fatty liver in both sexes. Fatty liver adjusted for the other factors was thus a risk factor for impaired fasting glucose and/or type 2 diabetes mellitus in both sexes (men odds ratio [OR] 1.91, 95% confidence interval [CI] 1.56–2.34 and women OR 2.15, 95% CI 1.53–3.01). The impact of fatty liver was stronger among the participants with a lower body mass index (men OR 0.92, 95% CI 0.86–0.99 and women OR 0.90, 95% CI 0.81–0.99, for one increment of body mass index).
Conclusion: Fatty liver is an independent risk factor for impaired fasting glucose and type 2 diabetes mellitus, having a stronger impact in those Japanese with a lower body mass index undergoing a health checkup.
Accumulation of triglycerides in hepatocytes is increasing due to consumption of a high-fat and high-calorie diet and a sedentary lifestyle and the prevalence of fatty liver is now 20–30% in Japan and other countries.1–7 Fatty liver is asymptomatic and the most common condition assessed by ultrasonography at health checkups.2,4,7,8 In particular, non-alcoholic fatty liver disease (NAFLD) is considered a hepatic consequence of the metabolic syndrome, closely associated with insulin resistance.7–11
It is widely accepted that impaired fasting glucose (IFG), elevated systolic blood pressure, a high body mass index (BMI), a family history of diabetes mellitus (DM), and adiposity and visceral fat distribution are risk factors for type 2 diabetes mellitus (T2DM).12–14 In addition, markers of liver injury may be associated with the metabolic syndrome and be independent predictors of T2DM.15–19 Thus, elevation of liver enzymes caused by fatty liver appears associated with insulin resistance.12,16,17,20 Although one study of Japanese men demonstrated that fatty liver assessed by ultrasonography was not a risk factor for T2DM,1 the majority of investigations have revealed a link between NAFLD and impaired glucose metabolism as well as diabetes.2,3,21,22 Recently, it was also reported that fatty liver was an independent risk factor for T2DM in participants including alcohol drinkers at a health checkup in Korea.4 Although it thus appears likely that fatty liver is a risk factor for T2DM, one study was performed in a cross-sectional manner2 and the others featured only small numbers of participants, only men or analysis of men and women together.3,4,21 Because sex and weight status may modify the relationship between metabolic risk factors and NAFLD,23 the sexes should be treated separately. Furthermore, because insulin resistance and hyperinsulinemia may be closely associated with NAFLD in the subjects with normal bodyweight and that non-obese subjects with NAFLD are prone to cardiovascular disease,24–26 it is important to determine the interaction between fatty liver and BMI regarding the risk of IFG and/or T2DM.
The metabolic syndrome is characterized by visceral adiposity (large waist circumference), dyslipidemia, hypertension, and IFG (≥ 110 mg). IFG itself is independently associated with cardiovascular risk factors such as hypertension and dyslipidemia as well as coronary artery calcification, subclinical atherosclerosis.27,28 There is also an independent link with T2DM.14,27,29 Therefore, it may be more beneficial to predict IFG, a prediabetic status, rather than T2DM itself in consideration of preventive measures against cardiovascular disease.
Therefore, in the present longitudinal investigation we assessed risk factors including fatty liver assessed by ultrasonography in 2000 for IFG or T2DM in both sexes of Japanese subjects undergoing a health checkup. Adjustment was made for age, BMI, elevated blood pressure or hypertension, family history of DM, alcohol drinking and smoking. A particular focus was on the relationship between fatty liver and BMI.
Design of the study
This study included retrospective longitudinal analyses to investigate whether fatty liver, assessed by ultrasonography, is associated with IFG or T2DM in apparently healthy Japanese subjects undergoing a health checkup. Informed consent was obtained from all participants.
Subjects of the study
The numbers of participants undergoing medical checkups, including ultrasonography in 2000 and 2005 were 26 247 (14 627 men and 11 620 women) and 32 548 (17 207 men and 15 341 women), respectively. A total of 14 617 (8377 men and 6240 women) underwent health checkups at both time-points. After exclusion of participants who had past and present illness of DM (551) and hepatic diseases (632), positive results for hepatitis viruses (159), fasting hyperglycemia in 2000 (1505), a total of 12 375 participants (men 6799, 49.2 ± 10.5 years old and women 5576, 50.6 ± 9.3 years old) were included.
Subjects provided data for family history of DM, alcohol drinking habits and smoking status through a self-administered questionnaire which was checked during individual interview by expert nurses in the center. Alcohol drinking habits were classified into occasional and daily. Family history of DM was defined if a parent had either a past history or present illness.
Age was categorized into four categories. Bodyweight was measured, in light clothing, to the nearest 0.1 kg and height to the nearest 0.1 cm. BMI was calculated as kg/m2 and divided into three categories according to the criteria determined by the Japan Society for the Study of Obesity.
Blood samples were taken from each participant after overnight fasting. Fasting blood glucose (FBG) was measured with Hitachi autoanalyzer models 7600 and 7700 (Hitachi Medical, Tokyo, Japan) and the presence of IFG or T2DM was defined as values between 110 and 125 mg/dL and 126 mg/dL or more, respectively.
Blood pressure was measured to the nearest 1 mmHg by an automatic sphygmomanometry (BP-203 RV III B; Nippon COLIN, Komaki, Japan). Elevated blood pressure or hypertension was diagnosed if resting blood pressures were 130/85 mmHg or more or if the participants had either a history of hypertension or use of antihypertensive medication, respectively.
Abdominal ultrasonographic examination was performed using convex-type real-time electronic scanners (SSA 250 and 300; Toshiba Medical, Tokyo, Japan) by 10 technicians without any information about any present illness. All images were printed on the sonographic papers and reviewed by other technicians and physicians. Fatty liver was assessed according to the modified criteria reported previously.30–33 Liver brightness (diagnosed by difference of more than 10 from the average of liver and renal cortical echo amplitudes), attenuation of echo penetration and decreased visualization of veins were included as criteria.
Logistic regression analyses were, respectively, performed to determine the risk of IFG or T2DM in both men and women separately. We evaluated two models in both sexes; an age-adjusted and a multivariate model with adjustment for age (< 40, 40–49, 50–59 and ≥ 60 years), BMI (< 25 kg/m2, 25–29.9 kg/m2 and ≥ 30 kg/m2), alcohol drinking (none, occasional, daily or unknown), smoking (never, ever or unknown), family history of DM (yes, no or unknown) and fatty liver (yes or no) which were assessed in 2000.
We also determined the interaction between fatty liver and BMI in a separate study. BMI was incorporated into the models as a continuous variable. In order to simplify interpretation, BMI was transformed by subtracting 22 (centerization). Statistical differences among groups were identified using one-way anova, followed by multiple comparisons using Bonferroni's method. The χ2-test and Fisher's test were employed for comparison of prevalence of fatty liver, IFG, and T2DM. Logistic regression analyses were performed using computer software (SPSS ver. 13.0 for Windows; SPSS, Chicago, IL, USA). P-values less than 0.05 were considered significant.
Incidences of newly diagnosed IFG and T2DM between 2000 and 2005 were, respectively, 5.9% and 0.8% overall (7.6% and 1.0% in men and 3.8% and 0.5% in women). They were 10.6% and 2.9% in men with fatty liver, and 5.2% and 0.6% in men without fatty liver. For women, the respective figures were 9.4% and 2.0% with fatty liver, and 2.6% and 0.4% without fatty liver. In both sexes, the differences were significant. The 78.0% of male and 71.3% of female participants with fatty liver in 2000 were assessed as fatty liver in 2005.
Table 1 shows the characteristics of the subjects by fatty liver status in men and women. BMI, systolic and diastolic pressures, and FBG were significantly elevated in the participants with fatty liver than those without fatty liver in both sexes. The prevalence of family history of T2DM was also significant.
Table 1. Clinical characteristics of the participants with and without fatty liver in 2000
|Men|| || || |
| Age||49.2 ± 10.5||49.5 ± 10.7||48.1 ± 9.6*|
| BMI (kg/m2)||22.9 ± 2.8||22.4 ± 2.5||25.3 ± 2.7*|
| Systolic blood pressure (mmHg)||118.2 ± 16.7||117.0 ± 16.6||122.8 ± 16.3*|
| Diastolic blood pressure (mmHg)||74.1 ± 10.9||73.4 ± 10.9||77.1 ± 10.6*|
| FBG (mg/dL)||95.1 ± 6.9||94.6 ± 6.9||97.1 ± 6.4*|
| Family history of diabetes mellitus||12.5||12.0||14.7*|
| Smoker (%)||45.7||47.2||40.1|
| Drinker (%)|| || || |
|Women|| || || |
| Age||50.6 ± 9.3||50.3 ± 9.3||53.7 ± 8.8*|
| BMI (kg/m2)||22.1 ± 2.8||21.8 ± 2.6||25.2 ± 3.0*|
| Systolic blood pressure (mmHg)||115.3 ± 17.0||114.4 ± 16.7||124.5 ± 17.2*|
| Diastolic blood pressure (mmHg)||71.0 ± 10.8||70.5 ± 10.6||76.2 ± 11.1*|
| FBG (mg/dL)||91.6 ± 7.2||91.2 ± 7.0||95.5 ± 7.1*|
| Family history of diabetes mellitus||15.3||14.8||19.8*|
| Smoker (%)||7.1||7.3||5.5|
| Drinker (%)|| || || |
Table 2 shows the age-adjusted and multivariate odds ratios with underlying fatty liver for IFG and T2DM. After adjustment for the potential confounders, fatty liver was a significant risk factor for IFG and T2DM in both men and women. The impact did not differ with the sex. The odds ratios (OR) were significantly larger among those with lower BMI. We thus found significant decrease of OR with fatty liver for IFG and T2DM, that is 0.92 (95% confidence interval [CI] 0.86–0.99) in men and 0.90 (95% CI 0.81–0.99) in women, for one increment of BMI.
Table 2. Multiple logistic regression analysis for impaired fasting glucose (IFG) or type 2 diabetes mellitus (T2DM)
|Men|| || || || |
| No fatty liver||1.00||Reference||1.00||Reference|
| Fatty liver||2.34||1.95–2.81||1.91||1.56–2.34|
|Women|| || || || |
| No fatty liver||1.00||Reference||1.00||Reference|
| Fatty liver||3.45||2.54–4.67||2.15||1.53–3.01|
The present study demonstrated that fatty liver as assessed by ultrasonography is an independent risk factor for IFG and T2DM in Japanese subjects undergoing health checkups. The incidence of newly diagnosed IFG or T2DM over the 5-year period was significantly higher in the participants with fatty liver than without fatty liver in both sexes. In addition, a significant interaction between fatty liver and BMI was observed and risk was higher among the leaner participants.
It has been demonstrated that fasting hyperglycemia, systolic blood pressure, BMI, family history of DM and visceral adiposity are risk factors for T2DM.12–14 Elevation of liver enzymes, including γ-glutamyltransferase and alanine aminotransferase is associated with the metabolic syndrome and is an independent predictor of T2DM.15–18 In most cases, elevation is due to fatty liver.12,16,17,20 Indeed, it has been shown that NAFLD is a risk factor for impaired glucose metabolism and T2DM,2–4,21 as confirmed for both sexes in the present study. It is well established that obesity is a strong risk factor for T2DM and a link has been found with increased BMI even within non-obese levels.34 Insulin resistance and hyperinsulinemia appear closely associated with NAFLD in the subjects with normal bodyweight24–26 and there may be increased risk of cardiovascular diseases.26,35 Indeed, we demonstrated herein that the impact of fatty liver on the risk factor of IFG or T2DM was stronger in leaner participants of both sexes. Taken together with the previous reports, we conclude that non-obese participants with fatty liver should be advised to make appropriate lifestyle changes.
The mechanisms by which fatty liver might lead to IFG or T2DM could not be elucidated in the present study. However, it is widely accepted that there is a close association with insulin resistance.7–10,20 Hepatic lipid accumulation causes impaired insulin clearance and defects in insulin suppression of glucose production which results in increased fasting serum glucose.11,13,36,37 On the other hand, it was demonstrated that percent bodyfat was an independent predictor of fatty liver in non-alcoholic and non-overweight subjects,38 suggesting that increased percent bodyfat may reflect central bodyfat distribution. Thus, we speculate that increased central bodyfat distribution may be related to the mechanism of a stronger impact of fatty liver on the leaner participants.
A major limitation of the present study was the retrospective longitudinal design. The subjects were limited to the Japanese participants undergoing voluntary health checkups at our center and might not necessarily be representative of the general population. Only 55.7% of the participants in 2000 received the health checkup in 2005. Although histological diagnosis would have been more accurate, liver biopsy is not an option at a health checkup. Therefore, we had to rely on ultrasonography for the purposes of the present study. However, this approach has been widely used as a non-invasive procedure with relatively high sensitivity and specificity for screening purposes5–7,30 and the 23.3% in men and 9.8% in women found in the present study are consistent with values in the previous reports.2,3,7,38 Finally, it is possible that misdiagnosis of IFG or T2DM have occurred in some cases because we had to rely on a single result of FBG for assessment.
In conclusion, fatty liver as assessed by ultrasonography may predict the development of IFG and T2DM in Japanese undergoing a health checkup, having strongest impact on those with a lower BMI. We propose that irrespective of BMI, the participants with fatty liver at health checkups should be advised to take action to reduce its risk factors to avoid possible development of diabetes. Cohort studies are now necessary to confirm the present findings.