Author(s): John Paul Verderese 1 , Zobair Younossi [*] 2
diabetes; nonalcoholic fatty liver disease; prediabetes
Methods & results
The subjects of this study were participants of the Ragama Health Study, a random cohort of urban Sri Lankan adults between the ages of 35 and 64 years of age. In 2007, 2984 members of this cohort were evaluated and grouped based on the presence or absence of nonalcoholic fatty liver disease (NAFLD) and diabetes. Subjects were considered to have NAFLD if they had fatty changes on liver ultrasound that were not attributed to excessive alcohol intake or underlying viral hepatitis (926 subjects with NAFLD, 1954 without NAFLD and 10 were excluded for excessive alcohol use), and were considered to be diabetic if they had a history of diabetes or fasting blood sugar of >125 mg/day (678 with DM, 308 of those with NAFLD; 2204 without DM, 618 of those with NAFLD).
Nondiabetic subjects with NAFLD (n: 618) were compared with nondiabetic subjects without NAFLD (n: 1586) in a prospective fashion for 3 years and subsequently assessed for incident diabetes. Baseline characteristics that were collected include age, BMI, mean waist circumference, gender, family history of diabetes, elevation of alanine transaminase (ALT) (>80 IU/l), impaired fasting glucose (IFG; fasting blood sugar: 100-125 mg/dl), comorbid hypertension (blood pressure: [greater than or equal]140/90) and comorbid dyslipidemia (low density lipoprotein: [greater than or equal]160 mg/dl, triglycerides: [greater than or equal]200 mg/dl and HDL: [greater than or equal]40 mg/dl). The two groups were slightly different as subjects with NAFLD had a higher baseline BMI, waist circumference, family history of diabetes, elevation in ALT, presence of IFG and comorbid conditions such as hypertension and hyplerlipidemia. Similar to baseline, incident diabetes was defined as a fasting blood sugar level of >125 mg/dl.
After multivariate analysis of those in the two groups who were not lost to follow-up (1914 out of the initial 2276 subjects) and accounting for subjects with interval development of unsafe alcohol consumption, the authors concluded that among the additionally collected baseline characteristics, independent predictors of incident diabetes were presence of IFG (hazard ratio [HR]: 4.8), NAFLD (HR: 1.6) and dyslipidemia (HR: 1.4). Most importantly, NAFLD was the only characteristic that predicted incident diabetes in those with and without baseline IFG.
Type 2 diabetes and its sequelae have a tremendous socioeconomic impact across the world. The financial burden...