The study, published in Frontiers in Gastroenterology, took an unusual two-pronged approach known as triangulation, pairing a Mendelian randomization analysis of large genetic datasets with a retrospective clinical study of patients whose portal pressure was measured directly. The logic behind triangulation is simple but powerful: no single observational study can fully rule out confounding, so researchers seek multiple lines of evidence that rely on different assumptions. If independent methods converge on the same answer, the conclusion becomes far harder to dismiss as an artifact of bias.

Mendelian randomization exploits a quirk of biology: genes are randomly allocated at conception and remain fixed throughout life. By using genetic variants that predict dietary fat intake as instruments, researchers can approximate a randomized trial of lifelong eating habits without actually assigning diets. Because the genes are inherited randomly, they are largely immune to the confounding factors that plague ordinary nutrition studies, such as obesity, insulin resistance, or the tendency of advanced liver disease itself to change what people eat. The method rests on three assumptions: the genetic variants must be strongly associated with the exposure, must not be linked to confounders, and must influence the outcome only through the exposure.

In practice, the team drew on genome-wide association study data covering 268,922 people of European ancestry, selecting 18 independent single nucleotide polymorphisms associated with fat intake as a percentage of total energy consumption. These instruments were strong by conventional standards, with a mean F-statistic of 34.2 against the threshold of 10 usually required, though together they explained only 2.8 percent of the variation in dietary fat. The primary analysis, an inverse variance weighted method, found that each standard deviation increase in genetically predicted fat intake was associated with 28 percent higher odds of portal hypertension, an odds ratio of 1.28 with a 95 percent confidence interval of 1.15 to 1.42 and a p-value of 1.0 times ten to the minus five.

Crucially, the association held up across every complementary method the team applied. The weighted median estimator, which remains valid even if up to half of the instruments are faulty, produced an odds ratio of 1.23. MR-Egger regression, designed to detect directional pleiotropy, where variants affect the outcome through pathways other than the exposure, yielded 1.35, and its intercept test showed no evidence of such pleiotropy. Cochran’s Q statistic found no significant heterogeneity among the SNP estimates, and the MR-PRESSO outlier-detection procedure identified no pleiotropic outliers, with the outlier-corrected estimate virtually identical to the original. A leave-one-out analysis confirmed that no single genetic variant was driving the result, and a Steiger directionality test verified that the causal arrow runs from diet to disease rather than the reverse. An independent FinnGen analysis of nearly 400,000 Finns, using liver cirrhosis as a surrogate endpoint with zero participant overlap, replicated the direction and significance of the effect.

The clinical arm of the study grounded these genetic findings in measurable physiology. The researchers retrospectively studied 72 patients with chronic liver disease who underwent hepatic venous pressure gradient, or HVPG, measurement at a tertiary hepatology center between 2021 and 2023. HVPG, the difference between wedged and free hepatic venous pressures, is the gold standard for assessing portal pressure, with values above 5 mmHg defining portal hypertension and values at or above 10 mmHg defining clinically significant disease. Dietary intake was assessed with a validated 168-item food frequency questionnaire administered by trained staff, and a high-fat diet score from 0 to 10 was calculated from consumption of red and processed meats, high-fat dairy, fried foods, cooking oils, and nuts and seeds.

The correlation was striking. The high-fat diet score correlated with HVPG at r = 0.700, and patients in the highest tertile of fat intake had a mean HVPG of 14.53 mmHg, compared with 12.29 mmHg in the moderate group and 9.37 mmHg in the low-fat group, a gap of roughly 5 mmHg between the extremes. The clinical stakes of that gap are considerable, because each 1 mmHg increase in HVPG is associated with approximately 10 to 15 percent higher risk of variceal bleeding, and even reductions of 2 to 3 mmHg are considered clinically meaningful. Nearly 94 percent of the high-fat diet group had clinically significant portal hypertension, versus only about 42 percent of the low-fat group.

Importantly, the association survived statistical adjustment. In a multivariate linear regression controlling for age, sex, BMI, bilirubin, albumin, and platelet count, each one-point increase in the high-fat diet score predicted a 0.85 mmHg increase in HVPG, and a partial correlation adjusting for BMI alone remained significant. That finding suggests dietary composition matters beyond its effect on body weight, consistent with emerging evidence that saturated and trans fats can exert direct hepatotoxic effects independent of overall adiposity. The regression model explained 62.3 percent of the variance in HVPG, and cross-validation and bootstrap correction confirmed it retained good predictive performance despite the modest sample size.

The authors stress that they did not directly test mechanisms, but several plausible biological pathways could connect fatty diets to raised portal pressure, all drawn from prior literature and acknowledged as speculative. Hepatic steatosis, the accumulation of fat inside liver cells, may increase intrahepatic vascular resistance by compressing sinusoids and reducing vascular compliance. Lipotoxicity from saturated fatty acids can trigger oxidative stress, endoplasmic reticulum stress, and activation of hepatic stellate cells, the cells that drive fibrosis. High-fat diets may also disrupt the gut microbiome and increase intestinal permeability, flooding the liver with bacterial endotoxin and fueling inflammation, while the associated insulin resistance can raise hepatic vascular tone by impairing nitric oxide signaling.

The study has real limitations, and the researchers are candid about them. Mendelian randomization captures the effect of lifelong genetic predisposition, which may not translate directly into the benefits of moderate dietary changes adopted in adulthood. The genetic analysis was restricted to people of European ancestry, the dietary data relied on self-reported questionnaires, and the clinical cohort was small and drawn from a single center with a retrospective design. The diet score also did not distinguish saturated from unsaturated fats. Even so, the convergence of genetic and clinical evidence on the same conclusion gives the finding unusual weight, and the authors argue it provides a rationale for testing dietary fat reduction as a potential adjunctive strategy for managing portal hypertension, pending confirmation in prospective intervention trials. For now, the message is a cautious one: what lands on the plate may help determine the pressure inside the portal vein.

Subject of Research: Causal relationship between high-fat diet and portal hypertension assessed by Mendelian randomization and clinical HVPG measurement

Article Title: High-fat diet and portal hypertension: a triangulation approach combining Mendelian randomization with retrospective clinical analysis

Article References: Xu, Y., Yang, S., Li, Y., Han, Q., & Wang, W. (2026). High-fat diet and portal hypertension: a triangulation approach combining Mendelian randomization with retrospective clinical analysis. Frontiers in Gastroenterology, 5, Article 1927449. https://doi.org/10.3389/fgstr.2026.1927449

Image Credits: AI Generated

DOI: 10.3389/fgstr.2026.1927449

Keywords: portal hypertension, high-fat diet, Mendelian randomization, hepatic venous pressure gradient, liver disease, causal inference, NAFLD, cirrhosis, GWAS, dietary fat, triangulation, hepatology