Genetics
Neither Lp(a) concentration nor apo(a) isoform size predicts new-onset diabetes in prediabetic adults, IT-DIAB study of 303 finds, challenging a general-population signal (Diabetes Metab 2025)
Original title: Lp(a) concentration and polymorphic size are not associated with new onset diabetes in individuals with prediabetes
This prospective IT-DIAB study followed 303 participants with impaired fasting glucose annually for 5 years to test whether low Lp(a), a signal previously reported in the general population, or Kringle-IV repeat polymorphism predicts new-onset diabetes (NOD, defined as fasting glucose 126 mg/dL or above) specifically in prediabetes. Of participants, 113 (37%) developed NOD during follow-up. Lp(a) concentration and Kringle-IV domain number did not differ significantly by NOD status, and the proportion with non-detectable (7 nmol/L or below) or elevated (above 125 nmol/L) Lp(a) was similar between those who did and did not develop NOD (68.1% vs. 63.7%, P = 0.46; 8.8% vs. 8.9%, P > 0.99). Kaplan-Meier and Cox models found no association between Lp(a) concentration or Kringle-IV domain number and NOD risk. The authors conclude that in this high-risk prediabetic population, neither Lp(a) concentration nor its genetically determined isoform size substantially predicts progression to type 2 diabetes.
Original abstract
Aim: Observational studies in the general population suggest that low concentrations of lipoprotein (a) [Lp(a)] are associated with an increased risk of type 2 diabetes. Here, we aim to determine whether Lp(a) plasma concentration and Kringle-IV (K-IV) repeat polymorphism were associated with new-onset diabetes (NOD) in individuals with prediabetes.
Methods: IT-DIAB is an observational, prospective study including 303 participants with impaired fasting glucose (fasting plasma glucose [FPG]: 110-125 mg/dl) followed annually for 5 years. The primary endpoint was the development of NOD, defined as a first FPG value ≥ 126 mg/dl during follow-up. Lp(a) concentrations were measured by immunoturbidimetry, apo(a) concentrations and the number of K-IV domains by mass spectrometry. Survival analyses for NOD were modeled using Kaplan-Meier curves and a multivariable Cox model, after binarization on threshold values of Lp(a) or K-IV.
Results: Among the participants, 113 (37%) developed NOD during follow-up. The concentrations of Lp(a) and the number of K-IV domains were not significantly different according to NOD status. Similarly, the percentage of patients with a non-detectable (≤ 7 nmol/l) or elevated (>125 nmol/l) Lp(a) concentration was similar between those with or without NOD: 68.1 vs 63.7% (P = 0.46) and 8.8 vs 8.9% (P > 0.99), respectively. Kaplan-Meier curves and Cox models did not show any association between Lp(a) concentration (threshold 7 nmol/l and 125 nmol/l) or number of K-IV domain (threshold 23) and the risk of NOD.
Conclusion: In a high-risk population, Lp(a) concentration or polymorphic size do not appear to be substantially associated with type 2 diabetes risk.
Summary written by lp-a.org from the published abstract; figures as published. Page updated 18 August 2026. Methods.