Subclinical cardiac involvement during diabetic ketoacidosis at type 1 diabetes onset in youth: biomarker elevation, altered diastolic filling, and reduced global longitudinal strain: a prospective cohort study

Figures

• 

• 

  Fig. 1. Study flow diagram. Flow chart of the study population showing enrollment and exclusion criteria for patients admitted with onset of diabetes mellitus from January 2024 to March 2025

• 

  Fig. 2. Troponin I and NT-proBNP at baseline and 24 h in DKA and Non-DKA groups. Box plots comparing cardiac biomarker levels between groups at baseline (A, B) and 24 h (C, D). Boxes represent the interquartile range (IQR) with medians indicated by horizontal lines. Whiskers indicate range excluding outliers. Extreme outliers are listed in the upper right corner of each panel. P values are shown in the upper left corner

• 

  Fig. 3. Longitudinal changes in troponin I and NT-proBNP in the DKA group. Troponin I and NT-proBNP levels in the DKA group are shown across four measurement time points (baseline, 24 h, 48 h, 72 h). Each data point represents an individual measurement. Violin plots illustrate the value distribution at each time point, with embedded box plots showing IQR and median. The bracket indicates a significant increase in NT-proBNP in the DKA group between baseline (n = 35) and 24 h (n = 43). Extreme outliers are listed in the upper right corner of each panel

• 

  Fig. 4. Echocardiographic parameters in DKA and non-DKA groups after 24 h. Statistically significant echocardiographic parameters are depicted for the DKA (n = 43) and non-DKA (n = 38) groups after 24 h. E/A ratios are shown on the left y-axis; peak a′-wave velocities (Med a′ Vmax, RV a′ Vmax) are plotted on the right y-axis. Each parameter is depicted by a distinct pictogram indicating the median; whiskers represent the interquartile range. Brackets show the p-values for comparison between DKA and non-DKA groups. MV E/A, mitral valve early-to-late diastolic filling ratio; e′/a′ medial, ratio of early to late diastolic tissue velocities at the septal mitral annulus; e′/a′ lateral, ratio at the lateral mitral annulus; RV e′/a′, right ventricular e′/a′ ratio; Med a′ Vmax, peak late diastolic tissue velocity at the septal mitral annulus; RV a′ Vmax, peak late diastolic velocity at the tricuspid annulus

• 

  Fig. 5. Left ventricular global longitudinal strain (GLS) by DKA status and timepoint. GLS values are shown for non-DKA patients after 24 h, and for DKA patients after 24 h and after 6 days. Boxes represent the IQR with medians indicated by horizontal lines and numerical values. Whiskers indicate range excluding outliers; dots represent outliers. Brackets indicate comparisons between groups and time points (significance threshold p < 0.05). Sample sizes were n = 16 (non-DKA, 24 h), n = 24 (DKA, 24 h), and n = 19 (DKA, 6 days). GLS was significantly less negative in the DKA group than in the non-DKA group at 24 h (p = 0.029), whereas the within-DKA comparison between 24 h and 6 days (paired Wilcoxon signed-rank test) was not significant. Less negative GLS values indicate worse systolic deformation. No 6-day measurement is available for the non-DKA group