The research, led by Guangpin Zeng of the People’s Hospital of Wenshan Prefecture together with Linran Zeng of Yuxi Zhongshan Hospital and Yinrong Zhu of the People’s Hospital of Wenshan, analyzed health-screening data from 6,998 adults collected in Wenshan between 2020 and 2024. The region was chosen deliberately: high-prevalence areas such as Southwest China offer the statistical power needed to detect even modest associations between infection status and metabolic markers. The cohort, drawn from a routine health-screening program, represents one of the larger single-population datasets assembled to examine this question, giving the findings unusual weight in a literature often dominated by small, heterogeneous studies.

At the center of the investigation was the triglyceride-glucose index, universally abbreviated as the TyG index. This simple mathematical measure—calculated as the natural logarithm of fasting triglycerides in milligrams per deciliter multiplied by fasting glucose in milligrams per deciliter, divided by two—has emerged in the past decade as a cheap, validated surrogate for insulin resistance. Unlike the hyperinsulinemic-euglycemic clamp, the labor-intensive gold standard that few clinical settings can justify for screening purposes, the TyG index requires nothing more than two values already present in virtually every routine blood panel. That accessibility has made it a favorite tool for large-scale epidemiological studies probing the metabolic consequences of infections, lifestyle factors, and chronic diseases.

Diagnosing H. pylori infection in the cohort relied on the carbon-14 urea breath test, a well-established technique that exploits the bacterium’s defining biochemical trick: its ability to produce urease, an enzyme that breaks down urea into carbon dioxide and ammonia. Participants ingested a small dose of carbon-14-labeled urea, and any H. pylori colonies in the stomach converted it into labeled carbon dioxide detectable in exhaled breath. A result exceeding 50 disintegrations per minute was classified as positive. The test is noninvasive, reasonably accurate, and widely deployed in Chinese screening programs, making it a practical choice for a study of this scale.

The raw numbers told what initially looked like an intriguing story. Of the 6,998 participants, 2,736—39.1 percent—tested positive for H. pylori, a prevalence consistent with the region’s reputation as a high-burden area. Those who carried the bacterium had a slightly but statistically significantly higher TyG index than their uninfected counterparts: 8.83 plus or minus 0.73 versus 8.76 plus or minus 0.70, a difference with a p-value below 0.001. In the univariate analysis, each unit increase in the TyG index was associated with roughly a 14 percent higher odds of infection, with an odds ratio of 1.14 and a 95 percent confidence interval of 1.06 to 1.21. On its face, the data appeared to support the idea that insulin resistance and H. pylori colonization travel together.

That impression evaporated under multivariable scrutiny. Once the researchers adjusted for age and sex—arguably the two most fundamental demographic confounders in any epidemiological analysis—the association collapsed. The adjusted odds ratio fell to 1.022, with a confidence interval spanning 0.950 to 1.100 and a p-value of 0.553, indistinguishable from no effect. Further statistical refinement, adding total bilirubin, alkaline phosphatase, gamma-glutamyl transferase, and low-density lipoprotein cholesterol to the model, did nothing to resurrect the signal. In the fully adjusted model, the only independent correlates of infection were age, male sex, and total bilirubin—demographic and biochemical characteristics, not metabolic state as captured by the TyG index.

The researchers went beyond conventional regression to stress-test their conclusions. Restricted cubic splines were used to model the relationship between the TyG index and infection flexibly, allowing the team to detect any nonlinear patterns that a simple linear term might miss. Stratified analyses examined whether the association might emerge only in particular subgroups—older participants, men, or people with specific liver enzyme profiles. Neither approach uncovered a hidden association. The pattern was consistent: whatever crude relationship existed in the unadjusted data was an artifact of demographic composition, not a genuine physiological link between the bacterium and insulin resistance.

Why would age and sex so thoroughly masquerade as a metabolic signal? The answer lies in the tangled web of correlations that characterizes observational health data. Older adults tend to have both higher rates of cumulative H. pylori exposure—having grown up in eras with poorer sanitation and higher transmission—and less favorable metabolic profiles. Men in this population, as in many others, showed different infection rates and different lipid and glucose levels than women. When such variables are left out of a statistical model, their effects bleed into whatever exposure-outcome relationship the model is testing, producing odds ratios that look meaningful but reflect demographic structure rather than biology. The Wenshan team’s findings suggest that earlier positive reports linking H. pylori to insulin resistance may have suffered from precisely this problem.

The authors are explicit about the practical implication: the TyG index should not be considered an independent marker or a screening tool for H. pylori infection, and clinicians have no basis for using metabolic panels to infer who might be carrying the bacterium. But the study’s larger contribution may be methodological. By demonstrating how dramatically a seemingly robust univariate association can dissolve under rigorous adjustment, the paper serves as a cautionary template for future research on infection-metabolism links. The team calls for prospective, longitudinal studies employing direct measures of insulin resistance—such as clamp techniques or validated surrogate indices tracked over time—to settle the question definitively, rather than relying on cross-sectional snapshots vulnerable to confounding.

The findings land at a moment of intense interest in the extragastric effects of H. pylori, with proposed connections ranging from cardiovascular disease to non-alcoholic fatty liver disease and neurodegeneration. Each of these hypotheses faces the same analytical gauntlet that the TyG question just failed to clear in this population: distinguishing true biological causation from the demographic and lifestyle factors that cluster around both infection and metabolic disease. For now, the message from Wenshan is one of disciplined skepticism. A bacterium that colonizes half of humanity certainly does something to the body—but according to this large, carefully adjusted analysis, distorting insulin sensitivity as measured by the TyG index does not appear to be among those things, at least not in the high-risk communities of Southwest China.

Subject of Research: The association between Helicobacter pylori infection and the triglyceride-glucose index as a marker of insulin resistance in a high-risk Chinese population

Article Title: Beyond the stomach: a cross-sectional analysis of H. pylori infection and the TyG index in a high-risk population from Southwest China

Article References: Zeng, G., Zeng, L., & Zhu, Y. (2026). Beyond the stomach: a cross-sectional analysis of H. pylori infection and the TyG index in a high-risk population from Southwest China. Gut Pathogens. https://doi.org/10.1186/s13099-026-00884-8

Image Credits: AI Generated

DOI: 10.1186/s13099-026-00884-8

Keywords: Helicobacter pylori, TyG index, insulin resistance, cross-sectional study, Southwest China, Gut Pathogens, metabolic syndrome, inflammation, urea breath test, confounding, epidemiology, gastroenterology