Serum Periostin as a Complementary Biomarker of Frailty in Type 2 Diabetes: A Pilot Study Using the FRAIL Scale

Figures

• 

  Figure 1. Scatter plots showing the correlation between the Frail scale and age (years), BMI (kg/m2), waist circumference (cm), diabetes duration (years), TBS, HDL-c (mg/dL), eGFR (mL/min/1.73 m2), and periostin (pmol/L) in patients with type 2 diabetes (n = 137). The associations were assessed using Spearman’s correlation coefficients (p < 0.05 for each association shown in the scatter plots). BMI: body mass index; TBS: trabecular bone score; HDL-c: high-density lipoprotein cholesterol; eGFR: estimated glomerular filtration rate.

• 

  Figure 2. Serum periostin levels in patients with type 2 diabetes according to FRAIL category. Violin plots show the distribution of serum periostin levels in the robust group (score = 0; n = 37), pre-frail group (score = 1–2; n = 74), and frail group (score = 3–5; n = 26). Boxplots represent the median and interquartile range, individual observations are shown as colored points, and the median value is additionally indicated for each group. Between-group differences were assessed using the Kruskal–Wallis test, with Holm-adjusted p values shown for statistically significant pairwise comparisons. p < 0.05.

• 

  Figure 3. ROC curves comparing the predictive performance of three logistic regression models for frailty. The blue curve corresponds to model_1, which includes all clinical variables associated with the FRAIL scale (age, BMI, waist circumference, duration of diabetes, cardiovascular disease, fractures, TBS, sedentary lifestyle, HDL-c, and eGFR); AUC = 0.858 (p < 0.001). The green curve represents model_2, which incorporates the same variables with the addition of serum periostin levels; AUC = 0.900 (p < 0.001). The red curve shows model_3, based solely on serum periostin; AUC = 0.710 (p < 0.001). Optimal cutoff points derived from Youden’s index are indicated by colored markers on each curve. ROC: receiver operating characteristic; AUC: area under the curve.