1. Overview of study populations.

	Intervention(s) and comparator(s)	Sample sizea	Screened/eligible [N]	Randomised [N]	Analysed [N]	Finishing trial [N]	Randomised finishing trial [%]	Follow‐up timeb
Umpierrez 2004a	I: s.c. insulin lispro	Arbitrary estimation of a difference between groups of ≥ 5 hours to determine ketoacidosis as being clinically important; a sample size of 20 participants was needed in each group to provide a power of 0.93, given an alpha level of 0.05, a SD of 4, and a 1:1 inclusion ratio	‐	20	20	20	100	Mean hospital stay: 4 days
	C: i.v. regular insulin			20	20	20	100
	total:			40	40	40	100
Umpierrez 2004b	I1: s.c. insulin aspart, every hour	Arbitrary estimation of a difference between groups of ≥ 4 hours to determine ketoacidosis as being clinically significant. A sample size of 15 participants was needed in each group to provide a power of 0.81, given an alpha error of 0.05 and a SD of 3	‐	15	15	15	100	Mean hospital stay: 3.4 days
	I2: s.c. insulin aspart, every 2 h			15	15	15	100	Mean hospital stay: 3.9 days
	C: i.v. regular insulin			15	15	15	100	Mean hospital stay: 4.5 days
	total:			45	45	45	100
Della Manna 2005	I: s.c. insulin lispro	‐	‐	25	25	25	100	Mean hospital stay: 2‐3 days
	C: i.v. regular insulin			21	21	21	100
	total:			46	46	46	100
Ersöz 2006	I: s.c. insulin lispro	‐	‐	10	10	10	100	‐
	C: i.v. regular insulin			10	10	10	100
		total:		20	20	20	100
Karoli 2011	I: s.c. insulin lispro	‐	‐	25	25	25	100	Mean hospital stay: 6 days
	C: i.v. regular insulin			25	25	25	100	Mean hospital stay: 6.6 days
			total:	50	50	50	100
Grand total	All interventions			110		110
	All comparators			91		91
	All interventions and comparators			201		201

^aAccording to power calculation in study publication or report
 ^bDuration of intervention and/or follow‐up under randomised conditions until end of study

‐ denotes not reported

C: comparator; I: intervention; i.v.: intravenous; s.c.: subcutaneous; SD: standard deviation