Table 5GRADE table summary for patients with ACS and hyperglycaemia and without a previous diagnosis of diabetes

Quality assessmentSummary of findings
No. of patientsEffectQuality
No. of studiesDesignLimitationsInconsistencyIndirectnessImprecisionOther considerationsIntensive insulinStandard therapyRelative risk (95% CI)Absolute
30-day mortality
1 (Weston et al. 2007)Observational studySeriousaNo serious inconsistencyNo serious indirectnessSeriousbNone116/841 (13.8%)327/1682 (19.4%)0.71 (0.58 to 0.86)6 fewer per 100 (from 3 fewer to 8 fewer)VERY LOW
30-day mortality
1 (Van der Horst et al. 2003)Randomised controlled trialNo serious limitationscSeriousdSeriouseSeriousbNone21/426 (4.9%)21/415 (5.1%)0.97 (0.52 to 1.81)0 fewer per 100 (from 5 fewer to −5 more)VERY LOW
7-day mortality (follow-up mean 7 days)
1 (Weston et al. 2007)Observational studySeriousaNo serious inconsistencyNo serious indirectnessSeriousbNone80/841 (9.5%)228/1682 (13.6%)0.70 (0.55 to 0.89)4 fewer per 100 (from 1 fewer to 6 fewer)VERY LOW
Inpatient heart failure
1(Cheung et al. 2006)Randomised controlled trialNo serious limitationscSeriousdNo serious indirectnessVery seriousgNone7/62 (11.3%)17/62 (27.4%)0.41 (0.18 to 0.92)16 fewer per 100 (from 2 fewer to 22 more)VERY LOW
Reinfarction (follow-up of up to 3 months)
2 (Cheung et al. 2006, Van der Horst et al. 2003)Randomised controlled trialNo serious limitationscSeriousdSerioush,iVery seriousgNone7/538 (1.3%)10/526 (2.1%)0.70 (0.27 to 1.82)1 fewer per 100 (from 1 fewer to 2 more)VERY LOW
Composite endpointj (follow-up mean 30 days)
1 (Van der Horst et al. 2003)Randomised controlled trialNo serious limitationscSeriousdSeriousiSeriousbNone38/476 (8%)46/464 (9.9%)adjusted RR 0.68k (0.44 to 1.05)3 fewer per 100 (from 6 fewer to 0 more)VERY LOW
Hypoglycaemial
1 (Van der Horst et al. 2003)Randomised controlled trialNo serious limitationscSeriousdSeriousiNo serious imprecisionNone0/4260/415No adverse effects were associated with intensive insulin therapyLOW
Subgroup analyses of mortality
1 (Van der Horst et al. 2003)Randomised controlled trialNo serious limitationscSeriousdSeriousiSeriousbNoneKillip class 1 (5/382)Killip class 1 (14/387)30-day mortality (RR 0.36, 95% CI 0.13 to 0.99, p = 0.05) was statistically significantly reduced by intensive insulin therapy in patients with Killip class 1. Mortality was not statistically significantly reduced in patients treated with intensive insulin therapy with Killip class 2 (RR 0.31, 95% CI 0.03 to 3.08, p = 0.32), Killip class 3 (RR 2.14, 95% CI 0.73 to 6.28, p = 0.17) and Killip class 4 (RR 1.45, 95% CI 0.51 to 4.13, p = 0.48).VERY LOW
Killip class 2 (1/21)Killip class 2 (2/13)
Killip class 3 (7/12)Killip class 3 (3/11)
Killip class 4 (8/11)Killip class 4 (2/4)
1 (Weston et al. 2007)Observational studySeriousaNo serious inconsistencyNo serious indirectnessSeriousbNoneSTEMI (80/509)

NSTEMI (57/359)
STEMI (193/755)

NSTEMI (196/1006)
30-day mortality was statistically significantly reduced in STEMI patients treated with intensive insulin therapy (RR 0.61, 95% CI 0.49 to 0.78, p < 0.0001). 30 day mortality was not statistically significantly reduced in NSTEMI patients treated with intensive insulin therapy (RR 0.81, 95% CI 0.62 to 1.07, p = 0.14). This was also reported at 7 days (STEMI RR 0.61, 95% CI 0.47 to 0.79, p = 0.0002, NSTEMI RR 0.76, 95% CI 0.53 to 1.08, p = 0.13).VERY LOW
Subgroup analyses of any composite endpointj
1 (Van der Horst et al. 2003)Randomised controlled trialNo serious limitationscSeriousdSeriousiSeriousbNone18/42636/430Composite endpoint (adjusted RR 0.47k, 95% CI 0.27 to 0.83, p = 0.01) was statistically significantly reduced by intensive insulin treatment in patients with Killip class 1.
Subgroup analyses of reinfarction
1 (Van der Horst et al. 2003)Randomised controlled trialNo serious limitationscSeriousdSeriousiSeriousbNone3/4266/430There was no statistically significant reduction in reinfarction in patients treated with intensive insulin therapy with Killip class 1 (adjusted RR 0.39k, 95% CI 0.09 to 1.63, p = 0.20)
a

There was no follow-up past the inpatient stay (outcome data was extracted from Office for National Statistics data using NHS numbers to identify patients). There were differences in the collection and/or recording of data across centres because blood glucose level and treatment strategy were not always available. There was also variation in what treatment was given.

b

95% CI includes both negligible effect and appreciable benefit and/or harm (defined as 25% relative risk reduction or relative risk increase).

c

The GDG considered downgrading based on the lack of blinding in this study; however, it was felt that it may not be feasible to conduct a blinded study in this situation.

d

Study not conducted in UK and practice may vary.

e

A median blood glucose of 8.5 mmol/litre was reported at admission, which the GDG felt may not be clinically indicative of hyperglycaemia and some patients without hyperglycaemia and a relatively low blood glucose would have been included.

g

This has been downgraded by two levels because of a small sample size, and the confidence interval includes both negligible effect and appreciable benefit and/or harm (defined as 25% relative risk reduction or relative risk increase).

h

The HI-5 study used glucose-insulin infusion for the intervention; the Van der Horst study used glucose-insulin-potassium infusion as the intervention.

i

The Van der Horst study included a small percentage of patients who had been diagnosed with diabetes for this outcome (approximately 10%). A median blood glucose of 8.5 mmol/litre was also reported in the Van der Horst study at admission, which the GDG felt may not be clinically indicative of hyperglycaemia and some patients without hyperglycaemia and a relatively low blood glucose would have been included.

j

Composite endpoints include death or recurrent infarction or repeat angioplasty.

k

Adjusted for age, gender, history, Killip class, infarct location and multivessel disease.

l

Cheung et al. 2006 only reported hypoglycaemia for all patients (diabetes and non-diabetes) and is not reported here.

NB: Adjusted relative risks are not shown for Weston et al. (2007) because figures reported in the paper were calculated using percentage dying in the untreated group divided by percentage dying in the insulin-treated group and were not consistent with reporting with the other papers.

Abreviations: 95% CI, 95% confidence interval; HR, hazard ratio; NSTEMI, non-ST-segment-elevation myocardial infarction; RR, relative risk; STEMI, ST-segment-elevation myocardial infarction.

From: 3, Evidence review and recommendations

Cover of Hyperglycaemia in Acute Coronary Syndromes
Hyperglycaemia in Acute Coronary Syndromes: Management of Hyperglycaemia in People with Acute Coronary Syndromes.
NICE Clinical Guidelines, No. 130.
Centre for Clinical Practice at NICE (UK).
Copyright © 2011, National Institute for Health and Clinical Excellence.

All rights reserved. This material may be freely reproduced for educational and not-for-profit purposes. No reproduction by or for commercial organisations, or for commercial purposes, is allowed without the express written permission of NICE.

PubMed Health. A service of the National Library of Medicine, National Institutes of Health.