Low lipoprotein(a) and high TyG index mark diabetes risk in prediabetes

Among people with prediabetes, low lipoprotein(a) combined with a high triglyceride-glucose index was associated with the highest observed risk of developing type 2 diabetes in a retrospective cohort study. During a median follow-up of seven years, 36 of 215 participants, or 16.7%, developed the disease. People who developed diabetes had higher triglyceride-glucose index values and lower lipoprotein(a) levels than those who did not.

The study examined 215 individuals with prediabetes, each of whom had at least five years of follow-up. The median observation period was seven years. The investigators focused on two measures: lipoprotein(a), abbreviated Lp(a), and the triglyceride-glucose index, or TyG, which the researchers describe as a surrogate marker of insulin resistance.

Receiver operating characteristic curve analysis was used to identify optimal thresholds for the two measures. Participants were then divided into four groups based on whether they had high or low Lp(a) and high or low TyG. The high-Lp(a)/low-TyG group served as the reference group.

Among participants who developed type 2 diabetes, the median TyG index was 9.0, with an interquartile range of 8.6 to 9.4. Among those who did not develop diabetes, the median was 8.6, with an interquartile range of 8.3 to 9.0. The difference was statistically significant at p less than 0.01.

The pattern for Lp(a) ran in the opposite direction: median Lp(a) was 12.0, with an interquartile range of 8.0 to 30.5, among participants who developed diabetes, compared with 19.0, with an interquartile range of 9.0 to 51.0, among those who did not; this difference had a p value below 0.05. In Cox regression analysis, the low-Lp(a)/high-TyG group had the highest observed risk of incident type 2 diabetes relative to the reference group. Its hazard ratio was 2.45, with a 95% confidence interval of 1.08 to 5.57, and p less than 0.05.

The other two groups did not show statistically significant associations. The low-Lp(a)/low-TyG group had a hazard ratio of 0.95, with a 95% confidence interval of 0.30 to 2.95 and p equal to 0.93, while the high-Lp(a)/high-TyG group had a hazard ratio of 1.22, with a 95% confidence interval of 0.70 to 2.11 and p equal to 0.48. High-sensitivity C-reactive protein, or Hs-CRP, was independently associated with incident type 2 diabetes in the regression analysis.

Its hazard ratio was 1.26, with a 95% confidence interval of 1.06 to 1.50, and p below 0.01. This result concerns an association within the study model; the supplied results do not establish that Hs-CRP causes diabetes or that it improves prediction beyond the other measures in practice. Together, the observations suggest that the combination of low Lp(a) and high TyG may identify a group with elevated risk among people with prediabetes.

The result is not evidence that either biomarker combination causes diabetes, nor does it show that the measures can predict an individual outcome with certainty. The study was retrospective, included a modest sample of 215 people with prediabetes, and reported that participants had at least five years of follow-up. The investigators describe the findings as hypothesis-generating and call for validation in larger studies to determine their potential value for risk stratification.

The confidence intervals, including the 95% interval of 1.08 to 5.57 for the low-Lp(a)/high-TyG association, indicate uncertainty around the estimated hazard ratio. Accordingly, the findings support further investigation of the combined signal, but the size and usefulness of that signal outside the studied population remain uncertain. The study was reported in the Journal of Translational Medicine as an open-access research article, with the publisher’s verified reference supplied separately.

Subject of Research: Association of lipoprotein(a) and triglyceride-glucose index with type 2 diabetes incidence in people with prediabetes

Article Title: Combined predictive value of lipoprotein(a) and triglyceride-glucose index for diabetes risk in a prediabetic population

Article References: Bosco, G., Guillerme, S., Bittar, R., Perregaux, J.-F., Redheuil, A., Ponnajah, M., De Francesco, E. M., Malaguarnera, R., Beaudeux, J. L., Le Goff, W., Scicali, R., Piro, S., & Gallo, A. (2026). Combined predictive value of lipoprotein(a) and triglyceride-glucose index for diabetes risk in a prediabetic population. Journal of Translational Medicine. https://doi.org/10.1186/s12967-026-09075-0

Image Credits: AI Generated

DOI: 10.1186/s12967-026-09075-0

Keywords: prediabetes, type 2 diabetes, lipoprotein(a), triglyceride-glucose index, TyG index, insulin resistance, diabetes risk, Cox regression, Hs-CRP, cohort study, Combined, predictive