Liver Enzyme Ratio Linked to Lower Hypertension Odds in Chinese Adults

A routine liver blood test that costs nothing extra to run may carry a hidden message about blood pressure. In a large cross-sectional analysis published in Frontiers in Public Health, researchers report that among 87,869 Chinese adults, a higher ratio of the liver enzymes aspartate aminotransferase (AST) to alanine aminotransferase (ALT) was independently associated with lower odds of hypertension. After adjusting for age, sex, and body mass index, each one-unit increase in the AST/ALT ratio corresponded to a 20 percent reduction in the odds of having elevated blood pressure, defined as systolic pressure of at least 140 mmHg or diastolic pressure of at least 90 mmHg. The association survived full statistical adjustment for metabolic markers and lifestyle factors, though the authors are careful to stress that the finding is observational, modest in magnitude, and no substitute for actually measuring blood pressure.

The scale of the underlying problem explains why the research team went looking for cheap auxiliary signals. Hypertension is among the most widespread chronic conditions worldwide and a principal modifiable driver of cardiovascular disease, stroke, kidney failure, and premature death. Recent global estimates indicate that more than 1.28 billion people aged 30 to 79 years are living with the condition, and the figure continues to climb. In China specifically, survey data show that 27.5 percent of adults, roughly 245 million people, are affected, while rates of awareness, treatment, and blood pressure control remain low. The authors argue that strategies enabling earlier detection in at-risk individuals are urgently needed, and that indicators already sitting in routine laboratory panels deserve scrutiny as potential flags for closer surveillance.

The biological rationale centers on insulin resistance, a metabolic state in which tissues respond poorly to insulin, forcing the pancreas to compensate with higher insulin output. This hyperinsulinemia can stimulate sympathetic nervous system activity and activate the renin-angiotensin-aldosterone system, enhance sodium retention by the kidneys, promote proliferation of vascular smooth muscle cells, and accelerate arterial stiffening, all of which push blood pressure upward. The reference standard for assessing insulin resistance, the hyperinsulinemic-euglycemic clamp, is costly and operationally demanding, so researchers have long sought simpler surrogate indices. Cross-sectional studies have already linked other surrogate measures, such as a triglyceride-glucose and waist circumference index, to hypertension in Indian adults, illustrating the appetite for practical proxies.

The AST/ALT ratio itself has a long pedigree. Originally described by Fernando De Ritis in 1959 as a tool for evaluating hepatic injury and classifying liver diseases, the ratio has more recently been tied to insulin resistance and a range of metabolic disorders. Data from the US National Health and Nutrition Examination Survey 2011 to 2016 showed that a reduced AST/ALT ratio was correlated with heightened insulin resistance among American adults, and in a Chinese cohort the reciprocal ratio ranked among the more reliable indicators of insulin resistance. Other investigations have connected a low ratio to incident type 2 diabetes, peripheral arterial disease, and cardiovascular mortality in hypertensive individuals. Direct evidence linking the ratio to hypertension, particularly in Chinese populations, had been scarce until now.

To fill that gap, the team conducted a secondary cross-sectional analysis of data hosted on the Dryad repository, originally collected through the Rich Healthcare Group health screening program across 32 sites in 11 Chinese cities spanning four major geographic regions. Of 211,833 eligible participants from the source cohort, the investigators required complete measurements of AST, ALT, systolic blood pressure, and diastolic blood pressure, sequentially excluding participants with missing values and one person with an implausible AST reading of zero, yielding a final analytic sample of 87,869. The group was 57.0 percent male with a mean age of 42.1 years, a mean BMI of 23.3 kilograms per square meter, and a mean AST/ALT ratio of 1.2. Missing covariate values were handled with multiple imputation using five replications.

The statistical strategy was deliberately layered. A crude model showed a marked inverse association, but the ratio was strongly patterned by demographic and metabolic factors, so the unadjusted estimate was heavily confounded. Model 1, adjusting only for age, sex, and BMI, produced an odds ratio of 0.80 per unit increase in the ratio. Adding fasting glucose, cholesterol fractions, and kidney function markers in Model 2 slightly attenuated the estimate to 0.84, and the fully adjusted Model 3, which further accounted for smoking and drinking, landed at 0.83. Because glucose and lipids may themselves lie downstream of the insulin resistance the ratio is thought to reflect, the authors present Model 1 as a co-primary estimate alongside the fully adjusted model, noting that the conclusion does not depend on whether those variables are treated as confounders or mediators. A sensitivity analysis removing kidney function markers left the association essentially unchanged.

Analyzed by quartiles, the pattern was consistent: compared with the lowest quartile, hypertension odds fell progressively across higher ratio categories, with fully adjusted odds ratios of 0.85 for the second quartile, 0.78 for the third, and 0.81 for the fourth, and a significant trend across all models. Smooth curve fitting then revealed something more interesting than a straight line. The dose-response relationship was nonlinear, with an inflection point at an AST/ALT ratio of approximately 1.010. Below that threshold, each unit increase in the ratio was associated with a 26.2 percent reduction in hypertension odds, whereas above it the inverse association weakened markedly. The likelihood ratio test confirmed the nonlinear structure, suggesting the ratio carries most of its information at lower values.

Stratified analyses uncovered clear effect modification. The inverse association was stronger among participants younger than 60 years, with an odds ratio of 0.81 compared with 0.89 in older participants, and both multiplicative and additive interaction tests were significant for age. It was also stronger among individuals with a BMI of 25 or higher, at 0.75 versus 0.88 in leaner participants, with the multiplicative interaction for BMI bordering on significance. The association held in both men and women, and no significant interactions emerged for smoking or drinking status. One compositional quirk deserves mention: 85.5 percent of the lowest ratio quartile was male compared with 30.0 percent of the highest, so the sex-adjusted estimates leaned heavily on within-sex contrasts, though the consistent within-sex associations argue against the overall finding being an artifact of that skew.

The biology behind the gradient appears to be driven mainly by ALT. In this cohort, the quartile differences were dominated by ALT, which ranged from a mean of 45.9 units per liter in the lowest ratio quartile to 11.6 in the highest, a roughly four-fold spread, while AST varied comparatively little. That pattern fits hepatic steatosis, in which the predominantly liver-cytosolic ALT rises while AST changes little, and a low ratio therefore primarily marks fatty liver and the associated insulin-resistant phenotype. An AST-driven elevation of the ratio, which typically reflects alcohol intake or muscle injury given AST’s expression in skeletal muscle, myocardium, and erythrocytes, is unlikely to explain the gradient here, since current drinkers made up only 1.1 percent of participants. Once insulin resistance is present, hyperinsulinemia can raise blood pressure through sympathetic and renin-angiotensin activation, renal sodium reabsorption, and vascular smooth muscle proliferation with collagen deposition and arterial stiffening.

The authors are explicit about the caveats. The cross-sectional design precludes causal inference, and reverse causation cannot be excluded, since hepatic steatosis can lower the ratio and promote hypertension through shared insulin-resistant pathways, meaning a low ratio may accompany rather than precede elevated blood pressure. Confounders such as physical activity, diet, antihypertensive medication use, and comorbid illness were unavailable for adjustment. The exposure was measured only at baseline, transient factors like recent alcohol intake, strenuous exercise, or muscle injury were not recorded, and the source population consisted of employees attending occupational health screening rather than a community-based random sample, raising the possibility of a healthy-worker effect. The inflection point of 1.010 should be regarded as cohort-specific until externally validated. The fully adjusted association is modest, and the team makes no claim of clinical screening value: at most, the ratio may serve as a zero-cost complementary flag within panels ordered anyway, and it cannot replace direct blood pressure measurement. Prospective studies are needed to determine whether the ratio predicts incident hypertension.

Subject of Research: Association between the AST/ALT ratio and hypertension in Chinese adults

Article Title: Association of aspartate aminotransferase to alanine aminotransferase ratio with hypertension in Chinese adults: a cross-sectional study

Article References: Zhang, X., Li, H., Liu, J., Ji, S., Hu, X., & Zhang, Z. (2026). Association of aspartate aminotransferase to alanine aminotransferase ratio with hypertension in Chinese adults: a cross-sectional study. Frontiers in Public Health, 14, Article 1936666. https://doi.org/10.3389/fpubh.2026.1936666

Image Credits: AI Generated

DOI: 10.3389/fpubh.2026.1936666

Keywords: AST/ALT ratio, hypertension, liver enzymes, insulin resistance, cross-sectional study, Chinese adults, hepatic steatosis, blood pressure, De Ritis ratio, public health, logistic regression, nonlinear dose-response