Van den Berghe, 2006

Web site designed and maintained by Chris Cairns © SICS EBM Group 2004

Intensive Insulin Therapy in Medical ICU patients

Intensive Insulin therapy reduces some markers of morbidity in medical ICU patients. It reduces mortality in patients admitted to ICU for longer than 3 days.

Level of Evidence: 1⁺(RCT with a low risk of bias)

Citation/s: Van den Berghe G, et al. Intensive Insulin Therapy in The Medical ICU. N Engl J Med 2006;354:449-61.
Lead author's name and fax: G Van den Berghe, greta.vandenberghe@med.kuleuven.be

Three-part Clinical Question: Does intensive insulin therapy improve outcome in medical ICU patients?

Search Terms: insulin, ICU, outcome, therapy

The Study: Single-blinded, randomised controlled trial with intention-to-treat.

The Study Patients: Adult ICU patients who, at admission to ICU, were predicted to require at least a three days of ICU care. Exclusions: surgical patients, medical patients able to receive oral nutrition (as unlikely to require 3 days of ICU), patients with DNAR orders. Mean APACHE 23(±10).

Control group (N = 605; 605 analysed): Standard care. Insulin started when BM>12mmol/l, maintained at 10-11mmol/l. When the BM fell below 10mmol/l the insulin was tapered off and stopped. When patients were haemodynamically stable, enteral feeding was started as early as possible as per the local regimen.

Experimental group (N = 595; 595 analysed): As control group in all aspects except for glycaemic control: Insulin was commenced when BM>6.1mmol/l and maintained between 4.4-6.1mmol/l.

The Evidence (All patients):

Outcome	Time to Outcome	CER	EER	RRR	ARR	NNT
Death	In hospital	0.400	0.373	7%	0.027	NS
Death	95% Confidence Intervals:			-7% to 21%	-0.028 to 0.082	NS

Non-Event Outcomes	Time to outcome/s	Control group		Experimental group	P-value
Newly acquired kidney injury (doubling of admission creatinine)	Hospital stay	8.9%	5.9%		0.04
Earlier weaning from mechanical ventilation	ICU discharge	Hazard ratio 1.15 1.02-1.44			0.03
Earlier discharge from ICU	ICU discharge	Hazard ratio 1.15 1.10-1.32			0.04

The Evidence (Patients in ICU for ³3 days):

Outcome	Time to Outcome	CER	EER	RRR	ARR	NNT
Death	Hospital discharge	0.525	0.430	18%	0.095	11
Death	95% Confidence Intervals:			5% to 32%	0.025 to 0.165	6 to 41
Death	90 days	0.491	0.422	14%	0.069	NS
Death	95% Confidence Intervals:			0% to 28%	-0.001 to 0.139	NS

Non-Event Outcomes

Time to outcome/s

Impact of insulin

P-value

Weaning from mechanical ventilation

ICU stay

Hazard ratio 1.43

(1.16-1.75)

<0.001

Discharge from ICU

ICU discharge

Hazard ratio 1.34

(1.12-1.61)

0.002

The Evidence (Patients in ICU for <3 days):

Outcome	Time to Outcome	CER	EER	RRR	ARR	NNH
Death	In-hospital	0.188	0.268	-43%	-0.080	12
Death	95% Confidence Intervals:			-85% to -1%	-0.159 to -0.001	902 to 6

Note the wide 95% CI’s

There was also a significant increased frequency of hypoglygaemic events in the intensive insulin group. These were not associated with any adverse outcomes.

EBM Comments:

1. Do the methods allow accurate testing of the hypothesis? Yes

2. Do the statistical tests correctly test the results to allow differentiation of statistically significant results? Yes

3. Are conclusions valid in light of the results? Yes

4. Did results get omitted, and why? No

5. Did they suggest areas of further research? Yes. A larger (5000 patient) multi-center study, adequately powered to detect a treatment effect for patients admitted for less than 3 days to ICU.

6. Did they make any recommendations based on the results and were they appropriate? No. As it is difficult to predict which patients will require ³3 ICU at the time of admission.

7. Is the study relevant to my clinical practice? Yes. Intensive glycaemic control is now widely applied in UK ICUs. This is as a result of this group’s previous study demonstrating benefit in surgical patients. In this study, the only sub-group to find a mortality benefit with intensive glycaemic control was those who were in ICU for at least 3 days. This group is difficult to predict at admission. For patients who stayed in the ICU for less than 3 days, intensive insulin therapy was associated with an increased mortality but also with decreased morbidity. Each unit will have to decide whether this study justifies tight gylcaemic control for all patients, at all stages of their ICU stay.

8. What level of evidence does this study represent? 1⁺

9. What grade of recommendation can I make on this result alone? B

10. What grade of recommendation can I make when this study is considered along with other available evidence? This is the only study examining the effect of intensive insulin therapy in the medical ICU patient population.

11. Should I change my practice because of these results? No. Each ICU will have to make a balanced decision on what policy to now adopt. Broadly speaking, until the results of larger multi-center studies are available, there are 3 options:

a. To ignore intensive insulin therapy.

b. To administer intensive insulin therapy to all patients on the strength of the current evidence.

c. To reserve tight glycaemic control for patients who have been in ICU for 3 or more days.

12. Should I audit my current practice because of these results? Yes

Appraised by: Dr Chris Cairns, Consultant, ICU, Stirling Royal Infirmary. ; 07 February 2006
Email: Chris.Cairns@fvah.scot.nhs.uk
Kill or Update By: February 2011

Citation: EBM Critical Appraisals. Scottish Intensive Care Society EBM Group. 2006 & JICS 2006 Vol7(1). Cairns CJS. Van den Berghe G, et al. Intensive Insulin Therapy in The Medical ICU. N Engl J Med 2006;354:449-61.

Reviewed & Edited by BC & BLT

Printer friendly view

Intensive Insulin Therapy in Medical ICU patients

NS