Triglyceride affects the association between estimated glomerular filtration rate and the onset of non-alcoholic fatty liver disease: A second analysis of a Chinese cohort study

OBJECTIVE: The role of triglyceride (TG) and estimated glomerular filtration rate (eGFR) effect modifiers on the risk of non-alcoholic fatty liver disease (NAFLD) is unknown. This study examined whether TG modifies the relationship between eGFR and incident NAFLD. METHODS: In a Chinese hospital from January 2010 to December 2014, 15,555 non-obese subjects were collected systematically for this retrospective cohort study. The target-independent and dependent variables were eGFR measured at baseline and NAFLD appearing during follow-up. The modified variable was TG measured at baseline. The multivariate Cox proportional hazards model was used to explore eGFR and TG’s association with NAFLD risk. We explored a priori interaction between eGFR and TG, and performed subgroup analyses to further assess whether the relationship between eGFR and incident NAFLD was modified by TG. We also explored the effect of TG and eGFR interaction on the risk of NAFLD. RESULTS: The mean age was 43.09 ± 14.92 years, and 8,131 (52.27%) were males. During a median follow-up time of 35.8 months, 2,077 (13.35%) individuals developed NAFLD. In the adjusted model, eGFR was negatively associated with incident NAFLD (HR = 0.984, 95% CI: 0.982, 0.987), while TG was positively related to NAFLD (HR = 1.582, 95% CI: 1.490, 1.681). TG could modify the relationship between eGFR and incident NAFLD. A stronger association between eGFR and NAFLD could be found in the participants without hypertriglyceridemia (HTG) (HR = 0.981, 95% CI: 0.978–0.984, P for interaction = 0.0139). In contrast, the weaker association was probed in the population with HTG (HR = 0.986, 95% CI: 0.983–0.989). At the same time, we also found an interaction between eGFR and TG in influencing NAFLD risk. In participants with decreased eGFR and HTG, the risk of NAFLD was significantly increased. Further, compared to non-HTG subjects with eGFR ≥ 116.56 ml/min/1.73 m(2), participants with HTG and eGFR < 82.88 ml/min/1.73 m(2) had about a fourfold increase in the risk (HR = 4.852 95% CI: 3.943–5.970) of NAFLD. CONCLUSION: eGFR and TG is independently associated with NAFLD risk. The association of eGFR with incident NAFLD is likely to be modified by TG in the Chinese non-obese population. There was an interactive effect between eGFR and TG in affecting NAFLD risk. In participants with decreased eGFR and hypertriglyceridemia, the risk of NAFLD is significantly increased.