Table 1 Overview of cohorts

Number of patients

1,660

1,172

19,738

5,529

4,562

3,601

5,032

2,290

1,440

Dates of admission to the ICU

1/09-12/09

10/09-3/11

7/07-6/10

4/09-3/12

9/05-12/10

4/08-6/11

10/05-6/11

3/10-5/12

2/01-3/09

Number and type of ICUs

Single 32-bed medical-surgical ICU of a university teaching hospital

Single 21-bed medical-surgical ICU of a university-affiliated teachinghospital

8 community-based hospitals, including 13 ICUs of mixed types, totaling227 beds

Single 82-bed mixed (medical, surgical, cardiac, neurosciences, trauma,burns, and transplant) ICU of a university teaching hospital

Single 18-bed medical-surgical ICU of a university-affiliated teachinghospital

Two medical-surgical ICUs (total 22 beds) of a university-affiliatedteaching hospital

Single 16-bed medical-surgical ICU of a university-affiliated teachinghospital

Single 10-bed surgical ICU of a university-affiliated teachinghospital

Single eight-bed medical ICU of a university hospital

Organizational details of centers

"Closed" format with intensivists supervising a team of critical carefellows, medical and surgical residents

Intensivist managed

All "Open" policy ICUs with mandate of critical care consult for allnon-pure cardiac admission

Intensivist managed

Intensivist managed

Intensivist managed

Intensivist managed, with medical and surgical residents

Intensivist managed, with medical and surgical residents

Medical intensivist managed, with medical residents

Glycemic targets

90-144 mg/dl

108-180 mg/dl

70-110 mg/dl from 1/20/05-10/1/2008 then

80-150 mg/dl up to 10/1/2011 then

100-160 mg/dl

<180 mg/dl

a. Prior to April 2009: 4.1-8.0 mM (73.9-144.1 mg/dl)

b. After April 2009: 7.1-10.0 mM (127.9-180.2 mg/dl)

<180 mg/dl

80-140 mg/dl from 10/1/05 to 1/10/07

80-125 mg/dl from 1/11/07 to 6/30/11

95-135 mg/dl since February 2002

<180 mg/dl to 06/03

80-110 mg/dl from 06/03-01/09

110-150 mg/dl from 01/09

Type of BG monitor

100% ABG analyzer (RapidLab 1200)

100% ABG analyzer

100% Accu-Chek Inform glucometers

100% ABG analyzer

100% ABG (Instrumentation Laboratory GEM 4000)

100% ABG analyzer

85% Accu-Chek Inform glucometers. 13% ABG analyzer

2% Central lab analyzer

98% Accu-check glucometer; 2% Central Lab analyzer

100% ABG analyzer

Source of blood

100% arterial

Venous or arterial blood

Capillary, venous, or arterial blood

98% arterial, 2% central venous

Arterial or venous blood

Venous or arterial blood

75% capillary

25% venous or arterial

70% Arterial, 23% central venous, and 2% capillary

100% arterial

Data acquisition

The blood glucose levels were extracted from the patient data-managementsystem (MetaVision, iMDsoft, Israel). Other patient data were extractedfrom the National Intensive Care Evaluation (NICE) database, maintainedby the NICE Foundation (reference: Arts D, de Keizer N, Scheffer GJ, deJonge E. Quality of data collected for severity-of-illness scores inthe Dutch National Intensive Care Evaluation (NICE) registry.Intensive Care Med 2002, 28:656-659.)

Glucose values captured automatically from arterial blood gas analyzerslinked to hospital information system. Demographic and clinical datamanually entered by trained data analysts into Australian National AdultIntensive Care database

ICUTracker Database linked to the hospital data systems

Glucose values captured automatically from arterial blood gas analyzerslinked to hospital information system. Demographic and clinical datamanually entered by trained data analysts into hospital database.

Glucose values captured automatically from arterial blood gas analyzerslinked to hospital information system Demographic and clinical datamanually entered by trained data analysts into Australian National AdultIntensive Care database

GAIA Database (Nihon Koden, Japan)

Comprehensive clinical database created in the ICU and linked to thehospital data systems

Glucostabilizer software program and ICUTracker Database.

Combination of clinical ward database (developed on the ICU) with BG-dataretrieved from the ABG analyzer