This was a cross-sectional analysis, carried out between August 2016 to February 2017, at a tertiary care centre in southern India. The study was carried out after an approval from the Institutional Ethics Committee (HDHCR/IEC/letter no: 2016143 dated 12^th June 2016). An informed consent was obtained from each participant before the study. The participants were ensured about the privacy and confidentiality of the exercise. The data of T2DM patients reporting into the diabetes clinic were screened for alcohol intake history.

Inclusion criteria: The patients with presence of T2DM for over six months and no history of alcohol intake as reported by the participant or the next of kin were included in the study.

Exclusion criteria: Patients with a history of jaundice, ascites, and features of liver cell failure, or those patients medicated with any drugs known to have hepatotoxic effects such as methotrexate, amiodarone, glucocorticoids, synthetic oestrogens, and nucleoside analogs were excluded from the study. Moreover, patients with chronic renal failure, chronic cardiac conditions of any nature, history of diabetic ketoacidosis, and major abdominal surgeries were not enrolled in the study. Finally, Hepatitis B surface Antigen (HBsAg) positive individuals were excluded from the study.

Sample size calculation: The prevalence of NAFLD was assumed to be 35% in nonalcoholic diabetic patients visiting the clinic over the period of six months (n=150) and a margin error of 5% was considered. A sample size of 100 was calculated at a 95% confidence interval (z score=1.96).

One hundred patients with no history of alcohol intake underwent physical examination, anthropomorphic measurements and laboratory tests in addition to an ultrasound of the abdomen to diagnose NAFLD. A detailed history was taken to elicit information on the known duration of diabetes, symptoms referable to liver disease, alcohol consumption, surgical, and drug history. A general physical examination was carried out that included the measurements of height, weight, Waist Circumference (WC), hip circumference, and examination of the abdomen for the liver, spleen for any free fluid in the abdomen. BMI was calculated in kg/m². WC was measured in standing position at the level of the umbilicus. Waist Hip Ratio (WHR) was calculated. Systolic and diastolic blood pressure was measured with a standard mercury manometer.

Venous blood was drawn after an overnight fasting for eight hours or more. Serum chemistry encompassing FBG, serum bilirubin, Serum Glutamic-Oxaloacetic Transaminase (SGOT), Serum Glutamic Pyruvic Transaminase (SGPT), Alkaline Phosphate (ALP), Serum Total Proteins (STP), and lipid profile were measured by standard laboratory procedures. HbA1c was also measured. PPBG was measured 2 hours after breakfast.

Hepatic ultrasonography was performed in all patients by a trained operator using a Logi tech EP5 with a 4 MHz probe. Hepatic steatosis was diagnosed as the presence of an ultrasonographic pattern consistent with “bright liver” with evident ultrasonographic contrast between hepatic and renal parenchyma, vessel blurring, focal sparing and narrowing of the lumen of the hepatic veins. Liver size >15 cm in the longitudinal plane was considered to represent hepatomegaly [12].

All the data were entered in Microsoft Excel 2016 and then exported to SPSS version 19 (Chicago, Illinois, USA). Statistical analysis was performed using student’s t-test. Significance was accepted at p<0.05.

Of the total of 100 T2DM patients 40 were male. The age of the study population ranged from 38 years to 75 years. (52.27±1.82 years). Their duration of diabetes ranged from 2 to 20 years (5.06±0.81 years). Among the 52 with NAFLD, 18 (34.60%) were men while 34 (65.40%) were women. There was no statistical significance with age and gender and the prevalence of NAFLD. The demographic and clinical parameters are shown in [Table/Fig-1].

Seventy-one percent (n=37) had no symptoms referable to the hepatobiliary system. Malaise was the most common symptom being present in 12 patients (23%), followed by right upper quadrant discomfort in nine patients (17.3%) the patients. 2 cases (3.8%) complained of abdominal distension. Jaundice was present in none. Hepatomegaly on clinical examination was present in 4 cases (7.7%). None had splenomegaly or free fluid in the abdomen.

Fatty liver alone was present in 38% (n=38), while it was present together with hepatomegaly in another 14% (n=14). Thus, the overall prevalence of fatty liver, i.e., NAFLD in the study population was 52%.

The duration of diabetes in diabetics with NAFLD ranged from 2 years to 15 years (5.13±1.30 years), while that in diabetics without NAFLD ranged from 3 years to 20 years. (4.98±0.79 years) (p=0.712).

The BMI of the diabetics with NAFLD was significantly higher than that of the patients without any signs of fatty liver disease. Similarly, overall obesity was prevalent in both the subgroups (13.46% vs 16.66%). The incidence of upper body obesity and the WHR was significantly higher in the NAFLD cohort.

The glycaemic control was better in patients without NAFLD than those with NAFLD. The fasting blood glucose in diabetics with NAFLD ranged from 90 mg% to 269 mg% (160.58±23.60 mg%), while that in diabetics without NAFLD ranged from 89mg% to 317 mg% (145.79±22.45 mg%) (p=0.0027).

The PPBG in diabetics with NAFLD ranged from 142 mg% to 402 mg% (218.88±38.08 mg%), while that in diabetics without NAFLD ranged from 122 mg% to 374 mg% (195.19±35.22 mg%) (p=0.0027). HbA1c was significantly higher in the NAFLD cohort.

The lipid profile amongst the two groups were also comparable. Serum cholesterol in diabetics with NAFLD ranged from 152 mg% to 363 mg%, (216.40±15.48 mg%), while that in diabetics without NAFLD ranged from 137 mg% to 324 mg% (198.5±19.18 mg%) (p<0.001).

The serum triglycerides in diabetics with NAFLD ranged from 67 mg% to 449 mg% (218.33±59.08 mg%), while that in diabetics without NAFLD ranged from 68 mg% to 358 mg% (170.04±52.82 mg%) (p<0.001).

There was no significant difference between the rest of the liver function tests serum bilirubin, SAP and STP amongst the two study cohorts. The duration of diabetes did not vary between the NAFLD and the non-NAFLD group.

The aim of this study was to determine the occurrence of NAFLD in the type 2 diabetics in a small tertiary centre setting with the help of simple clinical features and an ultrasound examination of the abdomen. In the Indian population, a high prevalence of all the components of metabolic syndrome in cases of NAFLD has been reported [13,14]. Several authors have reported the prevalence of NAFLD in the general population, in obese individuals and in those undergoing specialised surgeries in India [15,16]. However, there exists a paucity of data in terms of prevalence of NAFLD in the Indian diabetic cohort especially when NAFLD is viewed as a component of metabolic syndrome.

Matteoni CA et al., stated that the prevalence of NAFLD in diabetics ranged from 25 to 75% or even higher [17]. The prevalence of NAFLD in diabetics in the present study was 52%. This high occurrence indicates the importance of management and early evaluation of NAFLD in T2DM patients.

Of the 40 men, 18 (45%) had NAFLD; on the other hand, of the 60 women, 34 (56.67%) had NAFLD, a difference that was not statistically significant. However, among the 52 with NAFLD, 18 (34.60%) were men while 34 (65.40%) were women. Women far outnumbered men in a ratio of 1.9:1 in diabetics with NAFLD. However, in large meta-analyses globally, the prevalence of NAFLD was significantly higher in male T2DM patients than female T2DM patients [18,19]. Gender differences in the prevalence of NAFLD in T2DM patients could be attributed to the gender differences in hormone levels and lipid levels [18]. Clark JM et al., observed that men have a higher prevalence of NAFLD than women (5.7% vs 4.6%, respectively) however, there was no significant difference in either gender in IR as calculated by Homeostasis Model Assessment (HOMA) or exercise level in their study [20]. Conducting a HOMA to measure IR for all diabetics might not be feasible in tertiary care setups in India.

Seventy one percent of the subjects had no symptoms referable to the hepatobiliary system. Malaise was the most common symptom being present in 23%. As for physical signs, hepatomegaly was the only physical finding noted in the present study and it was present in 7.7% of the subjects studied. Several authors have also noted that an enlarged liver is the only sign noted at physical examination [21,22]. Given the insidious nature of the disease, it is not surprising to see malaise being the most common symptom in less than one-quarter of the study population. The most frequent symptoms are fatigue and right upper quadrant pain or dullness, although many patients have no symptoms [3]. Mild or moderate hepatomegaly is reported to be one of the most common physical examination findings [20-22].

The mean duration of diabetes in the NAFLD subgroup was 5.13±1.30 years (range 2-15 years) and did not differ significantly from those without NAFLD. A negative correlation between the prevalence of NAFLD and the duration of diabetes has been observed in several studies [23,24]. This negative correlation can be attributed to the greater degree of hyperinsulinemia in early T2DM driving the uptake of Free Fatty Acids (FFAs) by hepatocytes [24].

Prevalence of NAFLD in patients with obesity or T2DM can be as high as 80-90% [19]. In the current study, while overall obesity was prevalent similarly in both the subgroups (13.46% vs 16.66%), upper body obesity was present in a significantly higher percentage of diabetics with NAFLD. The WHR was higher in the NAFLD population with a statistical significance. Visceral fat has been considered as a direct predictor than subcutaneous fat for Nonalcoholic Steatohepatitis (NASH) and is related to the severity of NAFLD. WC is an alternative marker for the visceral fat and is strongly associated with triglyceride accumulation in liver cells, liver inflammation and, liver fibrosis [25].

The glycaemic control reflected by FBG, PPBG, and HbA1c was poorer in diabetics with NAFLD compared to diabetics without NAFLD. Similar observations have been made by other concomitant studies [26,27].

Similarly, total cholesterol, triglycerides, and LDL-cholesterol were significantly higher in diabetics with NAFLD compared to diabetics without NAFLD. Dyslipidaemia and hypercholesterolemia have been considered as direct predictors of NAFLD in several other studies [27-29]. In fact in recent times, NAFLD is considered as the hepatic manifestation of the metabolic syndrome.

Serum Bilirubin was within normal limits in diabetics with and without NAFLD and did not differ significantly between the two groups. The transaminases, even though within normal limits, were significantly higher in those with NAFLD compared to those without NAFLD. Systemic Arterial Pressure was within normal limits in both the subgroups and was similar. STPs, while being within normal limits, was lower in those with NAFLD. Mild elevations of liver enzymes in the upper normal range are associated with features of metabolic syndrome and NAFLD [30]. The high prevalence of metabolic syndrome and hyperlipidaemia in NAFLD patients suggests that risk stratification and aggressive treatment is needed to control the risk of cardiovascular diseases in these patients [1]. The typical biochemical pattern in hepatic steatosis due to NAFLD is, increased levels of transaminases, Alanine Aminotransferase (ALT) levels higher than Aspartate Aminotransferase (AST). This helps in differentiating between NAFLD from alcoholic liver injury, with the latter normally associated with a high AST:ALT ratio [31].

The study was conducted on a limited sample size of 100 patients. The second limitation is the self-reported alcohol consumption record during the initial screening.

It is concluded that NAFLD is common in T2DM being present in a little over a half of all diabetics and that poor glycaemic control, dyslipidaemia, and overall as well as regional adiposity are significantly associated with its development. Further, WC and WHR are easy to measure predictors of NAFLD. Ultrasound is a relatively inexpensive, non-invasive and accessible, compared to other diagnostic techniques, for screening for the presence of fatty liver in clinical settings and, especially in population studies. Therefore, plotting simple clinical features and a bedside USG assessment can provide a practical tool for precision medicine in diabetes so that physicians can tailor the management on the basis of clinical features of IR and NAFLD.

In India, there is a lack well designed prospective studies profiling large cohorts of patients with well-characterised NAFLD who are followed for long periods. Studies of this kind will provide the type of evidence needed to encourage clinicians, researchers, and health policy experts to focus on NAFLD as one of the most common chronic human diseases.