Louise J M Alferink1, Jessica C Kiefte-de Jong2,3, Nicole S Erler4, Robert J de Knegt1, Ewout J Hoorn5, M Arfan Ikram2, Harry L A Janssen1,6, Herold J Metselaar1, Oscar H Franco2,7, Sarwa Darwish Murad1. 1. Department of Gastroenterology and Hepatology, Erasmus MC, University Medical Center, CA Rotterdam, Netherlands. 2. Department of Epidemiology, Erasmus MC, University Medical Center, CA Rotterdam, Netherlands. 3. Department of Public Health and Primary Care, Leiden University Medical Center, LUMC-Campus, VA The Hague, Netherlands. 4. Department of Biostatistics, Erasmus MC, University Medical Center, CA Rotterdam, Netherlands. 5. Division of Nephrology and Transplantation, Department of Internal Medicine, Erasmus MC, CN Rotterdam, Netherlands. 6. Toronto Centre of Liver Disease, Toronto General Hospital, University Health Network, Toronto, Ontario, Canada. 7. Institute of Social and Preventive Medicine, University of Bern, Bern, Switzerland.
Abstract
OBJECTIVE: Our group recently showed that animal protein was independently associated with nonalcoholic fatty liver disease (NAFLD). We hypothesize that this may be explained by a high diet-dependent acid load [dietary acid load (DAL)]. METHODS: This cross-sectional study is embedded in a prospective population-based cohort. We estimated DAL proxies via food-frequency questionnaires using potential renal acid load (PRAL; using dietary protein, phosphorus, potassium, calcium, and magnesium intake), net endogenous acid production (NEAP; using protein and potassium intake), and the animal protein-to-potassium ratio (A:P). We defined NAFLD using ultrasound after excluding secondary steatogenic causes. We used logistic regression models-adjusted for sociodemographic, lifestyle, and metabolic traits-on categorized [quartile (Q)1 to 4] and continuous DAL proxies (allowing for nonlinearity) and NAFLD. RESULTS: We included 3882 participants, of which 1337 had NAFLD. All DAL proxies were higher, meaning more acidic, in individuals with NAFLD (PRAL, -2.9 vs -5.5 mEq/d; NEAP, 37.0 vs 35.1 mEq/d; and A:P, 13.3 vs 12.4; all P < 0.001). The highest Q of DAL proxies was associated with NAFLD independent of sociodemographic and lifestyle confounders, but significance dissipated after correction for metabolic confounders and multiple testing. However, the P value for nonlinearity was significant in all DAL proxies (P < 0.001). Natural cubic splines performed better with than without DAL proxies in the fully adjusted model (all P ≤ 0.038). The highest probability of NAFLD was found for an acidic diet. CONCLUSIONS: This study showed an independent nonlinear association between an acidic diet and NAFLD. Further studies with acid-base biomarkers are needed, but our findings might provide a mechanistic explanation for the harmful association between an animal protein-rich diet and NAFLD.
OBJECTIVE: Our group recently showed that animal protein was independently associated with nonalcoholic fatty liver disease (NAFLD). We hypothesize that this may be explained by a high diet-dependent acid load [dietary acid load (DAL)]. METHODS: This cross-sectional study is embedded in a prospective population-based cohort. We estimated DAL proxies via food-frequency questionnaires using potential renal acid load (PRAL; using dietary protein, phosphorus, potassium, calcium, and magnesium intake), net endogenous acid production (NEAP; using protein and potassium intake), and the animal protein-to-potassium ratio (A:P). We defined NAFLD using ultrasound after excluding secondary steatogenic causes. We used logistic regression models-adjusted for sociodemographic, lifestyle, and metabolic traits-on categorized [quartile (Q)1 to 4] and continuous DAL proxies (allowing for nonlinearity) and NAFLD. RESULTS: We included 3882 participants, of which 1337 had NAFLD. All DAL proxies were higher, meaning more acidic, in individuals with NAFLD (PRAL, -2.9 vs -5.5 mEq/d; NEAP, 37.0 vs 35.1 mEq/d; and A:P, 13.3 vs 12.4; all P < 0.001). The highest Q of DAL proxies was associated with NAFLD independent of sociodemographic and lifestyle confounders, but significance dissipated after correction for metabolic confounders and multiple testing. However, the P value for nonlinearity was significant in all DAL proxies (P < 0.001). Natural cubic splines performed better with than without DAL proxies in the fully adjusted model (all P ≤ 0.038). The highest probability of NAFLD was found for an acidic diet. CONCLUSIONS: This study showed an independent nonlinear association between an acidic diet and NAFLD. Further studies with acid-base biomarkers are needed, but our findings might provide a mechanistic explanation for the harmful association between an animal protein-rich diet and NAFLD.
Authors: Louise J M Alferink; Nicole S Erler; Robert J de Knegt; Harry L A Janssen; Herold J Metselaar; Sarwa Darwish Murad; Jessica C Kiefte-de Jong Journal: Eur J Epidemiol Date: 2020-04-22 Impact factor: 8.082