We read with great interest the article by Das et al.1 published in a recent issue of HEPATOLOGY. We agree with the authors that there is a significant prevalence of nonalcoholic fatty liver (NAFL) in the nonobese population and potentially significant liver disease related to NAFL in India. Even though our data were hospital-based, we suggested earlier that Indians with nonalcoholic fatty liver disease (NAFLD) do not suffer from the morbid overweight and obesity seen in the West but do have a metabolic profile (including insulin resistance) similar to that of their Western counterparts.2, 3 Even though the mean body mass index (BMI) was only 28.7 kg/m2, the majority of our patients suffered from overweight (21%), obesity (68%), or central obesity (82%) according to the Asia-Pacific guidelines.3 Although Das et al. used Asia-Pacific cutoffs for central obesity, their figures for overweight and obesity would have been higher (by at least 48%) even in the community population if they had used the Asia-Pacific cutoffs for defining overweight (BMI >23 but <25 kg/m2) and obesity (BMI ≥25 kg/m2).4, 5 Lower BMI cutoffs have been suggested for Asian populations because of the findings of higher percentages of body and visceral fat at a given BMI and higher morbidity and mortality rates at lower BMIs in Asians versus other populations.6, 7
Das et al.1 looked only at serum ferritin levels in their patients and found these to be normal; we studied in detail the serum (serum iron, transferrin saturation, serum ferritin, and total iron binding capacity), hepatic iron overload (Perls' Prussian blue staining), and hemochromatosis gene mutations and did not find these to be abnormal in Indian patients with NAFLD.8, 9 Similarly to Das et al., we also observed mild histological disease (a mild/moderate grade of inflammation with no or mild/moderate fibrosis) in patients with NAFLD presenting with elevated aminotransferase levels.2, 3 Only 51% of our patients (22 of 43) had histological nonalcoholic steatohepatitis (NASH), and none had cirrhosis at presentation,3 whereas for Das et al., 31% of the patients had NASH, and 2.4% of the patients (4) had cirrhosis due to NAFL.
Despite mild histological disease at presentation, patients with histological NASH have a propensity to progress to cirrhosis and hepatocellular carcinoma (HCC) and at that stage may be identified only by the presence of surrogate markers of NAFLD.10 We recently studied surrogate markers of NAFLD in 65 patients with cryptogenic cirrhosis and in 39 patients with cryptogenic hepatocellular carcinoma (CHCC) and compared the results to those for 50 patients with virus-related cirrhosis (VCC) and 39 patients with virus-related hepatocellular carcinoma (VHCC) of comparable age, gender, and liver disease severity. The study was approved by the institute's ethical review committee, and all patients provided informed consent. The diagnosis of cirrhosis was based on clinical findings, biochemistry, imaging, the demonstration of varices on gastroscopy, and histology (when available), and the presence of HCC was based on the practice guidelines of the American Association for the Study of Liver Diseases.11 All possible etiologies for cirrhosis and HCC, including viral etiologies, autoimmune etiologies, Wilson's disease, and iron overload, were excluded. Screening for occult hepatitis B virus infection (by total antibodies against core antigen) and celiac disease (by anti-tissue transglutaminase antibodies, anti-endomysial antibodies, and duodenal biopsy) was also performed in 16 and 10 patients, respectively. Abnormal metabolic parameters and metabolic syndrome were defined according to Adult Treatment Panel III criteria12 with a modified waist circumference for the Asia-Pacific region.5 The mean BMI was higher in patients with cryptogenic cirrhosis (26.06 ± 5.96 kg/m2) versus patients with VCC (22.12 ± 1.71 kg/m2, P = 0.0001). A higher number of patients with cryptogenic cirrhosis had an abnormal waist circumference [38 (58.5%) versus 15 (30%), P = 0.004], type 2 diabetes mellitus [26 (40%) versus 5 (10%), P = 0.0007], and lower serum high-density lipoprotein levels [35 (53.8%) versus 3 (6%), P = 0.0003] in comparison with patients with VCC. Patients with CHCC had a higher BMI (24.35 ± 4 versus 22.5 ± 3.4 kg/m2, P = 0.03) and a higher prevalence of type 2 diabetes mellitus [15 (38.5%) versus 7 (17.9%), P = 0.04] in comparison with patients with VHCC. There was no difference in abnormal high-density lipoprotein, serum triglycerides, or hypertension between patients with CHCC and patients with VHCC. The prevalence of metabolic syndrome was also similar in the two groups of patients with cirrhosis and HCC.
In conclusion, the higher prevalence of metabolic risk factors, if they are taken as surrogate markers of NAFL, suggests that NAFL is an important cause of both cryptogenic cirrhosis and CHCC and thus contributes to significant liver disease in India.