Urinary pteridines as a discriminator of atherosclerotic risk in patients with diabetes

Abstract

Background and aims: We have recently proposed urinary pteridine level as a useful biomarker of oxidative stress in a general population. However, the significance of urinary pteridines in patients with diabetes is unknown.

Methods: The relationships of the level of urinary pteridine derivatives with d-dimer, ankle-brachial pressure index (ABI), and known cardiovascular risk factors were investigated in patients with type 2 diabetes.

Results: Urinary pteridine level showed significant positive correlations with urinary15-isoprostane F_2t, female gender, history of smoking and d-dimer and significant inverse correlations with history of alcohol drinking, body mass index (BMI) and ABI. ABI was significantly lower and d-dimer was significantly higher in the highest tertile group of pteridines than in the lowest tertile group. The odds ratios of the highest vs. lowest tertiles for low ABI and high d-dimer were significantly higher than the reference level. The above relationships of urinary pteridines with ABI and d-dimer were not altered when age, gender, BMI, hemoglobin A1c and history of alcohol drinking were used as explanatory variables in multivariable analyses. History of smoking confounded the relation of pteridines with ABI but not that with d-dimer. However, in logistic regression analysis, the association between pteridines and ABI remained significant with adjustment for history of smoking.

Conclusion: Urinary pteridine level was associated with d-dimer and ABI, which reflect blood coagulability and arterial flow to the lower extremities, respectively, and is thus thought to be a useful discriminator of thromboatherosclerotic risk in patients with diabetes.

Keywords: Blood coagulation; Diabetes; Oxidative stress; Peripheral arterial disease; Pteridines.

Fig. 1

Histograms of urinary pteridines (A) and log-transformed pteridines (B).

Fig. 2

Correlations of log-transformed pteridines with log-transformed 15-isoprostane F_2t (A) and urinary creatinine (B). Pearson's correlation coefficients are given in the figures. Asterisks denote significant correlations (∗∗, p < 0.01).

Fig. 3

Comparison of mean levels of ABI (A) and log-transformed d-dimer (B) in the tertile groups of urinary pteridines. In multivariable analysis, age, gender, BMI, hemoglobin A1c and history of alcohol drinking were adjusted (Multivariable analysis 1 and 2). In addition, history of smoking was adjusted in Multivariable analysis 2. Asterisks denote significant differences from the 1st tertile for pteridine (∗, p < 0.05; ∗∗, p < 0.01). #, a marginally significant difference from the 1st tertile for pteridine (p = 0.077).

Fig. 4

ROC curves for urinary pteridine level to predict low ABI and high d-dimer. Cutoff points of urinary pteridine level, yielding maximal sensitivity plus specificity for predicting low ABI and high d-dimer, in the ROC curves were determined. The value with arrow, indicating the cutoff points in the figures, are cut-off values. Specificity and sensitivity of the cutoffs were 66.4% and 84.2%, respectively.