Stroke: GKI infusion no better than standard therapy
|
GKI infusion
in acute stroke does not lower plasma glucose better than standard therapy.
Increased
risk of biochemical, but not symptomatic hypoglycaemia.
No
difference in survival.
Level 1- evidence (RCT, but small
samples, one more event either way could have big impact, small samples
always limit our certainty of effects) |
Citation/s:
Scott JF, et al. Glucose Potassium Insulin Infusions in the Treatment of Acute
Stroke Patients With Mild to Moderate Hyperglycaemia: The Glucose Insulin in
Stroke Trial (GIST) Stroke 1999 30 (4); 793-9
Lead author's name and fax:
Prof CS Gray, University Dept of Medicine for the Elderly, Sunderland Royal
Hospital, Kayll Rd, Sunderland SR4 7TP c.s.gray@ncl.ac.uk
Three-part Clinical Question:
Among stroke patients with hyperglycaemia, does intensive insulin therapy
influence mortality?
Search Terms: Clinical Trials, Hyperglycaemia, Insulin and Stroke.
The Study:
The Study Patients: All patients >18years of age with a clinical
diagnosis of acute (within 24 hrs) stroke and a blood glucose level of 7 -17
mmol/L admitted to a DGH. Excluded if >24 hrs, CHF, Renal Failure, Anaemia,
pneumonia, GCS
< 4.
Control group (N = 25; 25 analysed): Standard stroke care plus 500ml
154mmol/L (Normal) saline at 100ml/hr through a peripheral vein in the
non-paretic arm for a minimum of 24 hours. All antihypertensive medication was
discontinued on admission.
Experimental group (N = 25; 25 analysed): Standard stroke care plus 500ml
10% Dextrose with 16U insulin & 20mmol KCl at 100ml/h through a peripheral vein
in the non-paretic arm for 24hrs. Monitoring & blood tests were identical for
each infusate. After the initial 24hrs, patients were maintained on the GKI, on
saline or on no fluids as clinical needs dictated. Clinical assessments were
undertaken by 1 of 2 observers at randomisation, at 24 hrs, 48 hrs, 7days & 4
weeks.
The Evidence:
|
Outcome |
Time to
Outcome |
CER |
EER |
RRR |
ARR |
NNT |
|
Mortality
|
4 weeks |
0.320 |
0.280 |
12% |
0.040 |
25 |
|
95% Confidence
Intervals: |
-67% to 92% |
-0.214 to
0.294 |
¥ = no effect |
|
Asymptomatic hypoglycaemia |
24 hours |
0 |
0.160 |
INF |
-0.160 |
-6 |
|
95% Confidence
Intervals: |
|
-0.304 to
-0.016 |
-61 to -3 |
|
Symptomatic hypoglycaemia |
24 hours |
0 |
0.040 |
INF |
-0.040 |
-25 |
|
95% Confidence
Intervals: |
|
-0.117 to
0.037 |
¥ = no effect |
Comments:
1. No clear benefit
from GKI in stroke and hyperglycaemia. No difference in blood sugar compared
to control group.
2.
No impact on clinically important end-points. Certainty is always
limited in small trials.
3. Bedside BM testing
corresponds closely to formal plasma samples.
4. Negative NNT
indicates increased risk of this outcome in GKI group.
Appraised by: Kevin Rooney, Dept of Anaesthesia & Intensive Care Medicine,
Royal Alexandra Hospital, Paisley; 30 December 2002 Email:
kd.rooney@virgin.net
Kill or Update By: December 2004
Printer friendly format
| 
Web site designed and maintained by
Chris Cairns
© SICS EBM Group 2004
