Feruze Yilmaz Enc1, Celal Ulasoglu1, Alev Bakir2, Yusuf Yilmaz3,4. 1. Department of Gastroenterology, Goztepe Teaching and Research Hospital, Medeniyet University. 2. Department of Medical Informatics and Biostatistics, Halic University. 3. Department of Gastroenterology, School of Medicine, Marmara University. 4. Liver Research Unit, Institute of Gastroenterology, Marmara University, Istanbul, Turkey.
Abstract
BACKGROUND: Higher hemoglobin levels have been associated with an increased risk for nonalcoholic fatty liver disease. Although the mechanism underlying this association is elusive, smoking has been previously related to both higher hemoglobin concentrations and an increased risk of fibrosis in nonalcoholic fatty liver disease. The present study was conducted to investigate formally the interaction among current smoking, hemoglobin levels, and risk for advanced fibrosis in patients with biopsy-proven nonalcoholic fatty liver disease. PATIENTS AND METHODS: We examined 433 Turkish patients with biopsy-proven nonalcoholic fatty liver disease. Advanced fibrosis (F ≥ 3) was identified on liver biopsy in 80 cases, whereas 84 patients were current smokers. Logistic regression models were used to evaluate the effect of current smoking on risk for advanced fibrosis, after adjusting for the effects of age, sex, BMI, diabetes, and metabolic syndrome. RESULTS: Preliminary analyses revealed the presence of substantial statistical interaction between current smoking and hemoglobin levels (P < 0.001). In separate multivariable analyses conducted in the entire cohort and in the subgroups of patients with high and low hemoglobin levels (according to median value in the study cohort: 14.4 g/l), current smoking was associated with increased risk for advanced fibrosis in patients with high hemoglobin (odds ratio: 3.32, 95% confidence interval: 1.23-7.21, P < 0.01) but neither in those with low hemoglobin (odds ratio: 0.71, 95% confidence interval: 0.28-1.81, P = 0.52) nor in the entire study cohort (odds ratio: 1.18, 95% confidence interval: 0.73-2.14, P = 0.79). CONCLUSION: Hemoglobin acts as a modifier in the association between current smoking and advanced fibrosis in nonalcoholic fatty liver disease.
BACKGROUND: Higher hemoglobin levels have been associated with an increased risk for nonalcoholic fatty liver disease. Although the mechanism underlying this association is elusive, smoking has been previously related to both higher hemoglobin concentrations and an increased risk of fibrosis in nonalcoholic fatty liver disease. The present study was conducted to investigate formally the interaction among current smoking, hemoglobin levels, and risk for advanced fibrosis in patients with biopsy-proven nonalcoholic fatty liver disease. PATIENTS AND METHODS: We examined 433 Turkish patients with biopsy-proven nonalcoholic fatty liver disease. Advanced fibrosis (F ≥ 3) was identified on liver biopsy in 80 cases, whereas 84 patients were current smokers. Logistic regression models were used to evaluate the effect of current smoking on risk for advanced fibrosis, after adjusting for the effects of age, sex, BMI, diabetes, and metabolic syndrome. RESULTS: Preliminary analyses revealed the presence of substantial statistical interaction between current smoking and hemoglobin levels (P < 0.001). In separate multivariable analyses conducted in the entire cohort and in the subgroups of patients with high and low hemoglobin levels (according to median value in the study cohort: 14.4 g/l), current smoking was associated with increased risk for advanced fibrosis in patients with high hemoglobin (odds ratio: 3.32, 95% confidence interval: 1.23-7.21, P < 0.01) but neither in those with low hemoglobin (odds ratio: 0.71, 95% confidence interval: 0.28-1.81, P = 0.52) nor in the entire study cohort (odds ratio: 1.18, 95% confidence interval: 0.73-2.14, P = 0.79). CONCLUSION: Hemoglobin acts as a modifier in the association between current smoking and advanced fibrosis in nonalcoholic fatty liver disease.