Glucose Measurements at Various Time Points During the OGTT and Their Role in Capturing Glucose Response Patterns

Adam Hulman; Róbert Wagner; Dorte Vistisen; Kristine Færch; Beverley Balkau; Melania Manco; Alain Golay; Hans-Ulrich Häring; Martin Heni; Andreas Fritsche; Daniel R. Witte

doi:10.2337/dc18-2397

e-Letters: Observations

Glucose Measurements at Various Time Points During the OGTT and Their Role in Capturing Glucose Response Patterns

Adam Hulman¹,²,³ ⇑,
Róbert Wagner⁴,⁵,⁶,
Dorte Vistisen⁷,
Kristine Færch⁷,
Beverley Balkau⁸,⁹,¹⁰,
Melania Manco¹¹,
Alain Golay¹²,
Hans-Ulrich Häring⁴,⁵,⁶,
Martin Heni⁴,⁵,⁶,
Andreas Fritsche⁴,⁵,⁶ and
Daniel R. Witte²,³

¹Steno Diabetes Center Aarhus, Aarhus University Hospital, Aarhus, Denmark
²Aarhus University, Aarhus, Denmark
³Danish Diabetes Academy, Odense, Denmark
⁴Division of Endocrinology, Diabetology, Nephrology, Vascular Disease, and Clinical Chemistry, Department of Internal Medicine IV, University Hospital of Tübingen, Tübingen, Germany
⁵Institute for Diabetes Research and Metabolic Diseases, Helmholtz Centre Munich, University of Tübingen, Tübingen, Germany
⁶German Center for Diabetes Research (DZD), Neuherberg, Germany
⁷Steno Diabetes Center Copenhagen, Gentofte, Denmark
⁸Centre for Research in Epidemiology and Population Health, University Paris-South, Paris, France
⁹Faculty of Medicine, University of Versailles-St. Quentin, Versailles, France
¹⁰INSERM U1018, University Paris-Saclay, Villejuif, France
¹¹Research Unit for Multi-factorial Diseases, Obesity and Diabetes, Istituto di Ricovero e Cura a Carattere Scientifico, Bambino Gesù Children's Hospital,Rome, Italy
¹²Division of Therapeutic Education for Chronic Diseases, University Hospitals of Geneva and University of Geneva, Geneva, Switzerland

Corresponding author: Adam Hulman, adam.hulman{at}ph.au.dk

Diabetes Care 2019 Apr; 42(4): e56-e57. https://doi.org/10.2337/dc18-2397

Article Figures & Tables

Tables

Table 1

Agreement (κ) between glucose pattern classification using five time points as reference and discrimination power (relative entropy) for different combinations of time points

Time points (min)					RISC cohort		TUEF cohort (external validation)
0	30	60	90	120	κ (95% CI)	Relative entropy	κ (95% CI)	Relative entropy
X				X	0.41 (0.37–0.45)	0.38	0.46 (0.43–0.49)	0.40
X	X				0.46 (0.42–0.49)	0.47	0.34 (0.31–0.36)	0.44
X		X			0.70 (0.67–0.73)	0.63	0.66 (0.64–0.68)	0.58
X	X			X	0.62 (0.58–0.65)	0.60	0.61 (0.59–0.63)	0.58
X	X	X			0.75 (0.72–0.77)	0.69	0.71 (0.69–0.73)	0.65
X		X		X	0.80 (0.77–0.83)	0.71	0.79 (0.77–0.81)	0.69
X	X	X		X	0.85 (0.82–0.87)	0.74	0.84 (0.82–0.85)	0.73
X	X	X	X	X	1 (reference)	0.79	1 (reference)	0.79

κ = 0 means agreement only by chance, while κ = 1 means perfect agreement. Relative entropy is a measure of adequate separation between classes taking values between 0 and 1. Higher relative entropy indicates that individuals can be classified with higher confidence (5).