We appreciate the interest of Dogru and colleagues in our recent article on the association between adiponectin serum levels and liver fibrosis in morbidly obese patients [1]. In their letter, they raise some important questions on our work that we are delighted to clarify. However, before we proceed, a word is in order regarding the main purpose of our study. This was conducted to explore the associations between plasma adiponectin concentrations and liver histology in morbidly obese patients and, for this reason, the main analysis related to patients with and without liver fibrosis within the group of obese subjects. It is important to note that the presence of fibrosis rather than inflammation is likely the major determinant of prognosis in nonalcoholic fatty liver disease (NAFLD) as shown in a recent publication [2] and that to non-invasively detect this feature is therefore clinically relevant. We included a control group in order to underscore the metabolic phenotype of morbid obese patients characterized by higher proportions of hypertension and type 2 diabetes mellitus as well as hyperleptinemia and hypodiponectinemia. Control subjects were selected from a population included in a previous epidemiological survey assessing the prevalence and natural history of cholelithiasis and NAFLD in Santiago, Chile [3,4]. For the purposes of the present study, controls were matched by sex and age and required to have a body mass index ≤25, serum alanine aminotransferase (ALT)≤ 30 IU/L in men and ALT≤19 IU/L in women, and an abdominal ultrasound without fatty liver. Based on their metabolic profile, these patients represent the opposite side of the metabolic syndrome spectrum. Therefore, nobody among them met the metabolic syndrome (MS) criteria and neither had an evidence of insulin resistance based on a HOMA-IR value >2.6 which is indicative of insulin resistance in non-diabetic subjects in a Chilean population according to Acosta et al. [5].In our study population, we did not perform glucose tolerance tests. Our patients were classified as diabetic or not from fasting glucose values as recommended. We agree that we cannot totally exclude impaired glucose-tolerant subjects in our study, but the presence of any impaired glucose-tolerant subjects in either of the two groups compared (obese patients with or without fibrosis) would have not changed the final results. Lower adiponectin serum levels were significantly associated with the presence of liver fibrosis and all the other variables highlighted by Dogru and colleagues such as body mass index, high density lipoprotein cholesterol and triglyceride levels, and also the proportions of subjects with hypertension and diabetes mellitus, although known to affect circulating adiponectin concentrations in other studies [6], did not show association with fibrosis in our study. In this regard, we are certain that multivariate analysis is the best way to assess a relationship between adipokines and hepatic fibrosis, in the setting of NAFLD, adjusted for major metabolic confoun...