Potential role of inflammation in relation to dietary sodium and β-carotene with non-alcoholic fatty liver disease: a mediation analysis

High sodium intake is a major risk factor for hypertension and has been linked to overweight/obesity and cardiovascular diseases. Prior work by the authors found a positive association between dietary sodium and NAFLD in the US population. NAFLD, considered a hepatic manifestation of metabolic syndrome, shares pathophysiology with hypertension and obesity, notably chronic inflammation. CRP is a systemic inflammation biomarker associated with disease and mortality risk, and RDW has been suggested as a marker of chronic inflammatory state. The authors hypothesized that high sodium intake increases NAFLD risk by upregulating inflammation and that higher dietary β-carotene intake attenuates sodium’s effect on NAFLD by downregulating inflammation. The study uses mediation analyses with NHANES 2007–2010 data to test these hypotheses.

The discussion references multiple observational studies reporting independent associations between higher sodium intake and NAFLD assessed by HSI and FLI in Korean cohorts and the PREVEND cohort, as well as prior NHANES analyses showing higher odds of NAFLD in the highest sodium intake quartiles. It also summarizes evidence that greater β-carotene or carotenoid levels are inversely associated with NAFLD prevalence and severity and may relate to NAFLD improvement over time. Mechanistic literature indicates high sodium can drive pro-inflammatory TH17 responses and hepatic steatosis/inflammation in animal models, whereas β-carotene and tomato-derived products may reduce hepatic inflammation and steatosis via anti-inflammatory pathways (e.g., SIRT1, adiponectin).

Data source and design: Cross-sectional analysis of NHANES 2007–2008 and 2009–2010. From 20,686 participants, adults ≥20 years with two reliable 24-h dietary recalls and valid energy intakes (women 500–5000 kcal/day; men 500–8000 kcal/day) were included. Exclusions: pregnancy, heavy alcohol use (>21 drinks/week men, >14 drinks/week women), positive HBsAg or anti-HCV, and missing key covariates. Final analytic samples: n=6725 for HSI-defined NAFLD; n=3237 subsample for FLI-defined NAFLD requiring fasting labs. Data collection: Demographics, socioeconomic status (family monthly poverty index categories ≤1.30, 1.31–1.85, >1.85), lifestyle factors (smoking status, drinking, sedentary time), anthropometrics (BMI, waist circumference), and dietary intake from two 24-h recalls averaged. Exposures: dietary sodium (mg/day) and β-carotene (μg/day). Dietary Inflammatory Index (DII) calculated per literature and used as a covariate. Laboratory measures: ALT, AST, GGT by enzymatic methods; HbA1c by HPLC; triglycerides enzymatic assay; CRP by latex-enhanced nephelometry; RDW from Coulter HMX. Diabetes defined as HbA1c ≥6.5% and/or glucose-lowering therapy. NAFLD definitions: HSI = 8 × (ALT/AST) + BMI (+2 if female; +2 if diabetes); NAFLD if HSI ≥36. FLI uses loge(TG), BMI, loge(GGT), waist circumference in standard formula; NAFLD if FLI ≥60. Statistical analysis: Sodium and β-carotene dichotomized by median and cross-combined into four groups (Low Na–High carotene; Low Na–Low carotene; High Na–High carotene; High Na–Low carotene). Group differences assessed by ANOVA/Chi-square. Odds ratios (ORs) for NAFLD estimated using logistic regression with multivariable adjustment for age, sex, ethnicity, education, family poverty level, smoking, drinking, sedentary time, total energy intake, DII, and serum creatinine. Dose-response evaluated using restricted cubic splines; sodium and β-carotene mutually adjusted when modeled separately. Mediation analysis adhered to AGREMA: sodium, CRP, and RDW standardized; path a (sodium→CRP or RDW) via linear regression; direct effect (sodium→NAFLD adjusting mediator) via logistic regression; indirect and total effects estimated using the R mediation package (causal approach). Analyses repeated with additional adjustment for β-carotene to observe its impact on mediation. Two-tailed P<0.05 considered significant; analyses in SAS 9.4 and R 3.6.3.

- Sample: 6725 adults (3119 men, 3606 women), mean age 53.2 ± 17.4 years; FLI subsample n=3237. - Group comparisons: Compared with high sodium–low carotene, the high sodium–high carotene group was older, more likely White, higher education and income, longer sedentary time, lower DII and log-CRP, more likely drinkers and less likely current smokers. - Odds of NAFLD by sodium/β-carotene combinations (Model 2 adjusted): HSI-defined NAFLD: High Na–High carotene vs High Na–Low carotene OR 0.84 (95% CI: 0.73–0.98). Low Na–High carotene vs reference OR 0.60 (0.51–0.71); Low Na–Low carotene OR 0.62 (0.53–0.73). FLI-defined NAFLD: High Na–High carotene vs reference OR 0.74 (0.60–0.92); Low Na–High carotene OR 0.53 (0.41–0.68); Low Na–Low carotene OR 0.58 (0.46–0.73). - Dose-response: Sodium positively associated with NAFLD (P<0.001) after adjustment including β-carotene; β-carotene inversely associated with NAFLD (P≤0.05) after adjustment including sodium. - Mediation by CRP and RDW (standardized effects): HSI-defined NAFLD: CRP indirect effect 0.0057 (95% CI: 0.0021–0.0091, P<0.0001); attenuated to 0.0050 (0.0021–0.0083, P<0.0001) after β-carotene adjustment. RDW indirect effect 0.0012 (0.0002–0.0027, P<0.0001); 0.0011 (0.0002–0.0022, P=0.040) after β-carotene adjustment. FLI-defined NAFLD: CRP indirect effect 0.0081 (0.0024–0.0162, P<0.0001); 0.0072 (0.0009–0.0121, P<0.0001) after β-carotene adjustment. RDW indirect effect 0.0009 (−0.0001 to 0.0027, P=0.080); 0.0008 (−0.0004 to 0.0019, P=0.200) after β-carotene adjustment. - Total and direct effects of sodium on NAFLD remained significant across models. Findings support partial mediation by inflammation (CRP, and to a lesser extent RDW) and attenuation of mediation after adjusting for β-carotene.

Findings support the hypothesis that higher dietary sodium is associated with increased odds of NAFLD and that this relationship is partially mediated by systemic inflammation, as reflected by CRP (and modestly by RDW for HSI). The observed attenuation of mediation after accounting for β-carotene intake is consistent with β-carotene’s anti-inflammatory properties and an associated reduction in NAFLD odds among participants with higher β-carotene intake even in the context of high sodium. These results align with prior epidemiologic studies showing positive associations between sodium intake and NAFLD and inverse associations between carotenoids and NAFLD. Mechanistic evidence from animal and human studies underscores sodium’s pro-inflammatory effects (e.g., TH17 induction) and β-carotene’s anti-inflammatory actions and potential to mitigate hepatic steatosis and inflammation. Overall, the mediation analyses provide human population-based evidence that inflammation is a plausible pathway linking sodium intake to NAFLD, and that higher β-carotene intake may blunt this pathway.

Higher sodium intake increases the odds of NAFLD, at least partly via upregulating inflammation. Dietary β-carotene appears to attenuate this association by downregulating inflammation. The study contributes mediation-based evidence for an inflammatory pathway between sodium and NAFLD and suggests that higher β-carotene intake may mitigate sodium-related hepatic steatosis. Future research should employ prospective designs with more objective sodium measures (e.g., multiple 24-h urinary sodium collections) and broader inflammatory biomarker panels to verify causality and elucidate mechanisms.

- Sodium intake assessed by two 24-h dietary recalls rather than multiple 24-h urinary sodium collections (gold standard). - NAFLD defined by HSI and FLI predictive equations rather than liver biopsy. - Inflammatory assessment limited to CRP and RDW; the inflammatory response is complex and may require additional biomarkers (e.g., IL-1β, IL-6, TNF-α, CD4+ T cell subsets). - Cross-sectional design with potential recall bias; causality cannot be established.