A machine-learning approach to predict postprandial hypoglycemia

Table 1 Clinical characteristics of enrolled study subjects

	Type-1 diabetes (55 three-day CGM datasets in 52 patients)	Type-2 diabetes (52 three-day CGM datasets in 52 patients)
Age (year)	40.0 (29.0-52.0)	63.5 (54.3-68.0)
Sex (male:female)	21:34*	21:31
Body weight (kg)	60.48 (52.35-69.41)	60.75 (54.60-70.37)
BMI (kg/ m²)	22.85 ±3.26	24.60 ±2.62
Duration of diabetes (years)	11.0 (6.0-18.0)	19.0 (13.3-25.0)
Insulin therapy (with insulin therapy: without insulin therapy)	55*:0	43:9
Insulin regimen basal:intermediate-acting: premix:MDI:CSII	3:1:6:44:1	20:3:11:9:0
Daily insulin dose (IU/day)	42.3 ±17.7	28.6 ±18.1
Daily insulin dose per body weight (IU/day/kg)	0.68 (0.53-0.82)	0.50 (0.30-0.60)
eGFR (ml/min/1.73 m²)	83.05 (71.98-96.95)	70.40 (51.30-82.50)
End stage renal disease [n (%)]	4 (7.3)	2 (3.8)
Liber cirrhosis [n (%)]	2 (3.6)	0 (0.0)
Heart failure with reduced ejection fraction [n (%)]	0 (0.0)	1 (1.9)
Pancreatic resection [n (%)]	2 (3.6)†	0 (0.0)
Acute infection [n (%)]	0 (0.0)	1 (1.9)
Pregnancy [n (%)]	1 (1.8)	0 (0.0)
Hemoglobin A1C (%)	7.94 ±1.13	8.31 ±1.32
C-peptide (ng/mL)	0.02 (0.02-0.15)	1.46 (0.80-2.44)

Continuous variables with normal distributions are expressed as mean ± standard deviation, whereas continuous variables with non-normal distributions were expressed as median (interquartile range)
^*Three female patients on insulin therapy were included twice because they participated twice
^†One of these two patients underwent total pancreatectomy; the other went through Whipple’s operation. Abbreviations: CGM, continuous glucose monitoring; BMI, body mass index; MDI, multiple daily injections; CSII, continuous subcutaneous insulin infusion; eGFR, estimated glomerular filtration rate

ISSN: 1472-6947