new england

The
journal of medicine
established in 1812 february 2, 2006 vol. 354 no. 5

Intensive Insulin Therapy in the Medical ICU

Greet Van den Berghe, M.D., Ph.D., Alexander Wilmer, M.D., Ph.D., Greet Hermans, M.D.,
Wouter Meersseman, M.D., Pieter J. Wouters, M.Sc., Ilse Milants, R.N., Eric Van Wijngaerden, M.D., Ph.D.,
Herman Bobbaers, M.D., Ph.D., and Roger Bouillon, M.D., Ph.D.

A bs t r ac t

Background
Intensive insulin therapy reduces morbidity and mortality in patients in surgical in- From the Departments of Intensive Care
tensive care units (ICUs), but its role in patients in medical ICUs is unknown. Medicine (G.V.B., P.J.W., I.M.) and Medi-
cal Intensive Care Medicine (A.W., G.H.,
W.M., E.V.W., H.B.) and the Laboratory for
Methods Experimental Medicine and Endocrinol-
In a prospective, randomized, controlled study of adult patients admitted to our ogy (R.B.), Catholic University of Leuven,
Leuven, Belgium. Address reprint requests
medical ICU, we studied patients who were considered to need intensive care for at to Dr. Van den Berghe at the Department
least three days. On admission, patients were randomly assigned to strict normal- of Intensive Care Medicine, Catholic Uni-
ization of blood glucose levels (80 to 110 mg per deciliter [4.4 to 6.1 mmol per liter]) versity of Leuven, B-3000 Leuven, Bel-
gium, or at greta.vandenberghe@med.
with the use of insulin infusion or to conventional therapy (insulin administered kuleuven.be.
when the blood glucose level exceeded 215 mg per deciliter [12 mmol per liter], with
the infusion tapered when the level fell below 180 mg per deciliter [10 mmol per N Engl J Med 2006;354:449-61.
Copyright © 2006 Massachusetts Medical Society.
liter]). There was a history of diabetes in 16.9 percent of the patients.

Results
In the intention-to-treat analysis of 1200 patients, intensive insulin therapy reduced
blood glucose levels but did not significantly reduce in-hospital mortality (40.0 per-
cent in the conventional-treatment group vs. 37.3 percent in the intensive-treatment
group, P = 0.33). However, morbidity was significantly reduced by the prevention of
newly acquired kidney injury, accelerated weaning from mechanical ventilation, and
accelerated discharge from the ICU and the hospital. Although length of stay in the
ICU could not be predicted on admission, among 433 patients who stayed in the ICU
for less than three days, mortality was greater among those receiving intensive in-
sulin therapy. In contrast, among 767 patients who stayed in the ICU for three or
more days, in-hospital mortality in the 386 who received intensive insulin therapy
was reduced from 52.5 to 43.0 percent (P = 0.009) and morbidity was also reduced.

Conclusions
Intensive insulin therapy significantly reduced morbidity but not mortality among
all patients in the medical ICU. Although the risk of subsequent death and disease
was reduced in patients treated for three or more days, these patients could not be
identified before therapy. Further studies are needed to confirm these preliminary
data. (ClinicalTrials.gov number, NCT00115479.)

n engl j med 354;5 www.nejm.org february 2, 2006 449

The New England Journal of Medicine

Downloaded from nejm.org on April 30, 2014. For personal use only. No other uses without permission.
Copyright © 2006 Massachusetts Medical Society. All rights reserved.
 The new england journal of medicine

H
yperglycemia and insulin resis- Study Design
tance are common in severe illness and On admission to the ICU, patients were randomly
are associated with adverse outcomes.1-4 assigned to receive either intensive insulin treat-
In a previous randomized, controlled study con- ment (intensive-treatment group) or conventional
ducted in a surgical intensive care unit (ICU), strict insulin treatment (conventional-treatment group).
control of blood glucose levels with insulin re- Treatment assignment was performed with the
duced morbidity and mortality,5,6 significantly re- use of sealed envelopes, stratified according to
ducing in-hospital mortality from 11 to 7 percent diagnostic category (Table 1), and balanced with
in the entire study population. In a subgroup of the use of permuted blocks of 10. In the conven-
patients who stayed in the ICU for three or more tional-treatment group, continuous insulin infu-
days, however, the benefit was much more pro- sion (50 IU of Actrapid HM [Novo Nordisk]) in
nounced, reducing mortality from 21 to 14 percent 50 ml of 0.9 percent sodium chloride) with the use
among patients treated for at least three days and of a pump (Perfusor-FM pump, B. Braun), was
from 26 to 17 percent among those treated for at started only when the blood glucose level exceed-
least five days. Complications, such as severe in- ed 215 mg per deciliter (12 mmol per liter) and
fections and organ failure, were reduced. Several was adjusted to maintain a blood glucose level of
potential mechanisms may explain these benefits between 180 and 200 mg per deciliter (10 and 11
— prevention of immune dysfunction,7 reduction mmol per liter). When the blood glucose level fell
of systemic inflammation,8 and protection of the below 180 mg per deciliter, the insulin infusion
endothelium9,10 and of mitochondrial ultrastruc- was tapered and eventually stopped.
ture and function.11 In the intensive-treatment group, insulin in-
It remains unclear whether intensive insulin fusion was started when the blood glucose level
therapy also improves the prognosis of patients in exceeded 110 mg per deciliter (6.1 mmol per liter)
a medical ICU, who often are more severely ill than and was adjusted to maintain normoglycemia (80
are patients in a surgical ICU and have a higher to 110 mg per deciliter [4.4 to 6.1 mmol per liter]).
risk of death.4,12,13 The study in a surgical ICU,5 The maximal continuous intravenous insulin
two studies of patients with diabetes with acute infusion was arbitrarily set at 50 IU per hour.
myocardial infarction,14,15 and observations in pa- At the patient’s discharge from intensive care,
tients with diabetes undergoing coronary-bypass a conventional approach was adopted (mainte-
surgery16 suggested that insulin-titrated blood glu- nance of blood glucose at 200 mg per deciliter
cose control should be continued for at least a or less).
few days to achieve a detectable outcome benefit. The dose of insulin was adjusted according to
We therefore conducted a randomized, controlled whole-blood glucose levels, measured at one-to-
study of patients in a medical ICU, targeting those four-hour intervals in arterial blood or, when an
requiring intensive care for at least a third day. arterial catheter was not available, in capillary
blood, with the use of a point-of-care glucometer
Me thods (HemoCue B-glucose analyzer, HemoCue). Adjust-
ments were made by the nurses in the ICU; the
Adult patients admitted to the medical ICU who usual number of nurses (2.5 full-time-equivalent
were assumed to require at least a third day of nurses per bed in the ICU) was not changed for
intensive care were eligible for inclusion. We ex- the study. The nurses used titration guidelines that
cluded surgical ICU patients and medical patients were adapted from the study in the surgical ICU.5
able to receive oral nutrition, because such patients When patients were hemodynamically stable,
usually need less than three days of intensive enteral feeding was started according to routine
care, and patients with do-not-resuscitate orders guidelines. The guidelines aimed at a total of 22
on admission (Fig. 1). Written informed consent to 30 kcal per kilogram of body weight per 24
was obtained from the closest family member, be- hours with balanced composition (0.08 to 0.25 g
cause patients were unable to give consent. The of nitrogen per kilogram of body weight per 24
protocol and consent forms were approved by hours and 20 to 40 percent of nonprotein kilo-
the institutional review board of the university. calories as lipids).17 Enteral feeding was attempt-
The study was carried out between March 2002 ed as early as possible.
and May 2005.

450 n engl j med 354;5 www.nejm.org february 2, 2006

The New England Journal of Medicine

Downloaded from nejm.org on April 30, 2014. For personal use only. No other uses without permission.
Copyright © 2006 Massachusetts Medical Society. All rights reserved.
 Intensive Insulin Ther apy in the Medical ICU

Data Collection
At baseline, data on demographic and clinical char-
2110 Evaluated
acteristics of the patients were obtained, includ-
ing information necessary to determine the se-
verity of illness and the use of intensive care
resources (Table 1). These data were scored ac- 863 Were excluded
387 Were expected to stay
cording to the Acute Physiology and Chronic in ICU <3 days (eating)
Health Evaluation (APACHE II)18 system and sim- 269 Had DNR orders on
admission
plified Therapeutic Intervention Scoring System- 150 Were postoperative
28 (TISS-28),19,20 with higher values indicating 38 Were participating in
another study
more severe illness and more therapeutic inter- 19 Were ineligible for
ventions, respectively. other reasons
47 Did not provide consent
Blood was systematically sampled and blood
glucose levels were measured on admission and
subsequently every four hours in all patients. More
frequent blood glucose measurements were per- 1200 Underwent randomization

formed whenever the attending nurse considered

them necessary and whenever there had been a
steep rise or fall in the blood glucose level on the
previous reading. Blood glucose levels that were
measured on admission and daily in the morn-
605 Assigned to conventional 595 Assigned to intensive
ing during the study, and hypoglycemic events treatment treatment
(defined as blood glucose levels of ≤40 mg per
deciliter [2.2 mmol per liter]) were analyzed.
According to clinical guidelines, blood cultures
were obtained whenever the central body tem- 381 Stayed in ICU for 3 days 386 Stayed in ICU for 3 days
perature exceeded 38.5°C or when other clinical
signs of sepsis were present.21,22 Results were in-
Figure 1. Patients in the Study.
terpreted by an investigator blinded to the treat-
All adult patients admitted to the medical intensive care unit (ICU) from
ment assignment. An episode of bacteremia was March 14, 2002, onward who were assumed to require at least a third day
defined by the first positive culture in a series. of intensive care were eligible for inclusion. Of those, 767 patients re-
To identify bacteremia with coagulase-negative mained in the ICU for at least a third day. DNR denotes do not resuscitate.
staphylococci, identical strains (compared by an-
tibiogram) in two or more positive blood cultures
were required.21,22 A distinction was made between creatinine twice that present on admission to the
primary and secondary bacteremia, depending on ICU 23 or a peak level of serum creatinine of >2.5 mg
whether or not a focus could be identified. per deciliter [220 μmol per liter]), days of inotro-
The clinical cause of a death in the ICU was pic or vasopressor support, presence or absence of
determined by a senior physician blinded to the hyperinflammation (defined as a C-reactive pro-
treatment assignments. The causes of deaths oc- tein level of >150 mg per deciliter), presence or
curring after discharge from the ICU could not be absence of bacteremia, prolonged (i.e., more than
identified. 10 days) use of antibiotics, and the presence or
absence of hyperbilirubinemia (defined as a biliru-
Outcome Measures bin level of >3 mg per deciliter [51 μmol per liter]).
The primary outcome measure was death from Use of intensive care resources was assessed on
any cause in the hospital. Secondary, predefined the basis of cumulative TISS-28 scores (the sum
outcome measures were mortality in the ICU, 90- of daily scores), indicating the total number of
day mortality, days to weaning from mechanical interventions per patient.19 We performed a pre-
ventilation, days in the ICU and in the hospital, defined subgroup analysis for patients staying in
the initiation of dialysis, new kidney injury during the ICU for at least a third day. A post hoc explor-
intensive care (defined as either a level of serum atory mortality analysis was performed censoring

n engl j med 354;5 www.nejm.org february 2, 2006 451

The New England Journal of Medicine

Downloaded from nejm.org on April 30, 2014. For personal use only. No other uses without permission.
Copyright © 2006 Massachusetts Medical Society. All rights reserved.
 The new england journal of medicine

Table 1. Baseline Characteristics of the Patients.*

Variable Intention-to-Treat Group Group in ICU for ≥3 Days

Conventional Intensive P Value Conventional Intensive P Value†
Treatment Treatment Treatment Treatment
(N = 605) (N = 595) (N = 381) (N = 386)
Male sex — no. (%) 382 (63.1) 356 (60) 0.24 243 (64) 224 (58) 0.10
Age — yr 64±16 63±16 0.61 64±16 62±16 0.20
BMI 24.8±5.1 25.1±5.5 0.29 24.6±5.1 25.4±5.9 0.06
Diagnostic category — no. (% of 0.99 0.70
patients in the category)
Respiratory 261 (51.0) 251 (49.0) 172 (47.9) 187 (52.1)
Gastrointestinal or liver 152 (49.7) 154 (50.3) 89 (55.6) 71 (44.4)
Hematologic or oncologic 51 (52.6) 46 (47.4) 37 (48.7) 39 (51.3)
Other sepsis 45 (50.0) 45 (50.0) 30 (46.9) 34 (53.1)
Cardiovascular 24 (48.0) 26 (52.0) 15 (45.5) 18 (54.5)
Neurologic 31 (50.8) 30 (49.2) 14 (50.0) 14 (50.0)
Renal 20 (45.5) 24 (54.5) 11 (45.8) 13 (54.2)
Metabolic 11 (55.0) 9 (45.0) 10 (66.7) 5 (33.3)
Other 10 (50.0) 10 (50.0) 3 (37.5) 5 (62.5)
History of cancer — no. (%) 128 (21.2) 134 (22.5) 0.57 90 (23.6) 98 (25.4) 0.57
Dialysis-dependent kidney failure 37 (6.1) 37 (6.2) 0.94 29 (7.6) 31 (8.0) 0.83
before acute event — no. (%)
Kidney failure on admission 120 (19.8) 119 (20.0) 0.94 92 (24.1) 82 (21.2) 0.34
to ICU — no.(%)‡
History of diabetes — no. (%) 97 (16.0) 106 (17.8) 0.41 58 (15.2) 59 (15.3) 0.98
Treated with insulin 51 (8.4) 65 (10.9) 27 (7.0) 41 (10.6)
Treated with oral antidiabetic 46 (7.6) 41 (6.9) 31 (8.1) 18 (4.7)
agent, diet, or both
Baseline APACHE II score§
Mean 23±9 23±10 0.50 24±9 24±10 0.95
>40 — no. (%) 18 (3.0) 26 (4.4) 0.19 11 (2.9) 18 (4.7) 0.19

patients for whom intensive care was limited or were transferred from intensive care had no access
who were withdrawn from intensive care by a se- to the results of blood glucose testing and were
nior attending physician within 72 hours after unaware of the study treatment assignment.
admission for reasons of futility.
To minimize the possibility of bias in assess- Statistical Analysis
ing the ICU stay caused by delays in the transfer On the basis of data from our previous study,5 we
of patients to a regular ward because of the un- hypothesized an absolute reduction in the risk of
availability of beds, patients were considered to death of 7 percent after at least three days of in-
be ready for discharge when they no longer needed tensive insulin therapy. Testing this hypothesis
vital-organ support and were receiving at least required a sample of 1200 patients for a two-sided
two thirds of their caloric intake by the normal alpha level of less than 0.05 and a beta level of 0.2
enteral route or when they were sent to a ward. in the targeted group of patients staying in the
Physicians on the general wards to which patients ICU for three or more days.

452 n engl j med 354;5 www.nejm.org february 2, 2006

The New England Journal of Medicine

Downloaded from nejm.org on April 30, 2014. For personal use only. No other uses without permission.
Copyright © 2006 Massachusetts Medical Society. All rights reserved.
 Intensive Insulin Ther apy in the Medical ICU

Table 1. (Continued.)

Variable Intention-to-Treat Group Group in ICU for ≥3 Days

Conventional Intensive P Value Conventional Intensive P Value†
Treatment Treatment Treatment Treatment
(N = 605) (N = 595) (N = 381) (N = 386)
Baseline TISS-28 score¶
Mean score 29±7 29±7 0.45 30±7 31±7 0.46
Score >33 — no. (%) 125 (20.6) 158 (26.6) 0.02 96 (25.2) 129 (33.4) 0.01
Blood glucose on admission 162±70 162±71 0.98 164±68 163±67 0.87
— mg/dl
Glycosylated hemoglobin on 6.2±0.9 6.3±0.9 0.12 6.2±0.9 6.3±0.9 0.21
admission — %∥
Plasma creatinine on admission
— mg/dl
Median 1.2 1.2 0.25 1.4 1.3 0.06
Interquartile range 0.9–2.1 0.8–2.1 0.9–2.4 0.8–2.1
Plasma urea on admission — mg/dl
Median 67 65 0.26 71 69 0.16
Interquartile range 40–110 36–104 45–115 37–106
Plasma ALT on admission — IU/liter
Median 29 30 0.50 33 31 0.55
Interquartile range 15–64 16–63 17–77 17–63
Plasma CRP on admission — mg/liter
Median 124 108 0.27 146 132 0.31
Interquartile range 39–226 36–218 55–236 48–229

* Plus–minus values are means ±SD. To convert values for glucose to millimoles per liter, multiply by 0.05551. To convert values for urea to
millimoles per liter, multiply by 0.357. ICU denotes intensive care unit, BMI body-mass index (the weight in kilograms divided by the square
of the height in meters), APACHE II Acute Physiology and Chronic Health Evaluation, TISS-28 Therapeutic Intervention Scoring System, ALT
alanine aminotransferase, and CRP C-reactive protein.
† P values for the comparison between the conventional-treatment group and the intensive-treatment group were calculated by Student’s
t-test, the Mann–Whitney U test, or the chi-square test, as appropriate.
‡ Established kidney failure on admission was defined as dependence on dialysis or a serum creatinine level >2.5 mg per deciliter. To convert
values for creatinine to micromoles per liter, multiply by 88.4.
§ Higher APACHE II scores indicate more severe illness, with a score greater than 40 representing the 90th percentile, indicating the most
severe illness.
¶ According to the TISS-28, each therapeutic intervention is assigned a number of points, with higher scores indicating a greater number of
therapeutic interventions. The sum of the points is calculated daily for each patient. A score greater than 33 is at the upper limit of the inter
quartile range.
∥ Glycosylated hemoglobin was measured by immunoturbidimetric assay (Dimension, Dade Behring) (normal range, 4 to 6 percent).

The baseline and outcome variables were com- for all well-known, clinically relevant baseline risk
pared with the use of Student’s t-test, the chi- factors. The effect on time to weaning from me-
square test, and the Mann–Whitney U test, as chanical ventilation and time to discharge from
appropriate. The effect of the intervention on time the ICU and from the hospital was assessed by
to death in the hospital was assessed with the use cumulative hazard estimates and proportional-
of Kaplan–Meier estimates and proportional-haz- hazards regression analysis, with censoring for
ards regression analysis. Patients discharged alive early deaths.
from the hospital were considered survivors. The The data are presented as means ±SD or me-
hazard ratios for death, calculated by propor- dians (with interquartile ranges), unless otherwise
tional-hazards regression analysis, were corrected indicated. Separate analyses were performed for

n engl j med 354;5 www.nejm.org february 2, 2006 453

The New England Journal of Medicine

Downloaded from nejm.org on April 30, 2014. For personal use only. No other uses without permission.
Copyright © 2006 Massachusetts Medical Society. All rights reserved.
 The new england journal of medicine

Conventional-treatment group Intensive-treatment group

Figure 2. Nutrition Administered to All 1200 Patients
during the First 14 Days of Intensive Care and Daily
A Morning Blood Glucose Levels during the First 10 Days
of Intensive Care.
2000
In Panel A, feeding at 0 represents the administration
Total Intake of Nonprotein Calories

of nutrition to patients admitted to the intensive care

1600 unit (ICU) after midnight between admission and 7 a.m.,
and 1 represents feeding on the first day after admis-
sion, from 7 a.m. on. Nutrition in the two groups was
(kcal/24 hr)

1200
similar. Total kilocalories are given as means ±SE. In
Panel B, in the box plot the fraction of nutrition admin-
800 istered by the enteral route is expressed as medians
(indicated by horizontal lines within the bars) and in-
terquartile ranges (with the 90th percentile indicated
400 by the I bar). In Panel C, among patients staying in the
ICU for three or more days, intensive insulin treatment
was continued until discharge from the ICU (mean,
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 12.5 days, with a range up to 65 days). P<0.001 for the
comparison between the two groups. To convert val-
Day
ues for glucose to millimoles per liter, multiply by
0.05551.
B
1.0
the intention-to-treat group and for the group stay-
Fraction of Kilocalories Administered

0.8 ing in the ICU for three or more days. For com-
parison with the results of our previous study in
by Enteral Route

the surgical ICU,5 the effects on patients in the

0.6
ICU for at least a fifth day were also documented.
P values were not adjusted for multiple compari-
0.4
sons. The study sponsors were not involved in the
design of the study, the collection, analysis, or in-
0.2 terpretation of the data, or the preparation of the
manuscript.
0.0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
R e sult s
Day
Nutrition and Blood Glucose Control
C Table 1 gives the baseline characteristics on ad-
180 P< 0.001
mission of all 1200 patients enrolled in the study,
160
including the 767 patients who stayed in the ICU
140 for at least a third day. Nutritional intake and blood
Blood Glucose (mg/dl)

120 glucose levels are shown in Figure 2 (for insulin

100 doses, see Table A in the Supplementary Appen-
dix, available with the full text of this article at
80
www.nejm.org). Hypoglycemia occurred more of-
60 ten in the intensive-treatment group than the con-
40 ventional-treatment group. Most patients who had
20 hypoglycemia had only one episode. The severity
of hypoglycemia was similar in the two groups
0
Admis- 1 2 3 4 5 6 7 8 9 10 (Table A in the Supplementary Appendix). No he-
sion modynamic deterioration, convulsions, or other
Day
No. of Patients events were noted in association with any hypo-
Conventional- 605 538 450 381 323 286 251 226 208 188 174 glycemic event. Mortality among patients in the
treatment group
Intensive-treat- 595 536 455 386 335 290 251 217 197 175 157 ICU who had hypoglycemia was 66.7 percent in
ment group the conventional-treatment group, as compared
with 46.4 percent in the intensive-treatment group

454 n engl j med 354;5 www.nejm.org february 2, 2006

The New England Journal of Medicine

Downloaded from nejm.org on April 30, 2014. For personal use only. No other uses without permission.
Copyright © 2006 Massachusetts Medical Society. All rights reserved.
 Intensive Insulin Ther apy in the Medical ICU

(P = 0.1); the in-hospital mortality was 73.3 per- fidence interval, 1.02 to 1.44; P = 0.03), along with
cent and 61.9 percent, respectively (P = 0.4). Two earlier discharge from the ICU (hazard ratio,
patients in the conventional-treatment group and 1.15; 95 percent confidence interval, 1.01 to 1.32;
three in the intensive-treatment group died with- P = 0.04) and from the hospital (hazard ratio,
in 24 hours after having a hypoglycemic event. 1.16; 95 percent confidence interval, 1.00 to 1.35;
Independent risk factors for hypoglycemia, aside P = 0.05) (Fig. 3). There was no significant effect
from intensive insulin therapy (odds ratio, 7.50; on bacteremia (reduction, 7 to 8 percent; P = 0.5),
95 percent confidence interval, 4.50 to 12.50; prolonged requirement of antibiotic agents (re-
P<0.001), were a stay in the ICU for three or more duction, 24 to 21 percent; P = 0.2), hyperbilirubi-
days (odds ratio, 3.33; 95 percent confidence in- nemia (reduction, 27 to 25 percent; P = 0.4), hyper-
terval, 1.95 to 5.70; P<0.001), renal failure requir- inflammation (reduction, 61 to 56 percent; P = 0.1),
ing dialysis (odds ratio, 1.94; 95 percent confi- or cumulative TISS-28 scores (reduction, 308±16
dence interval, 1.19 to 2.84; P = 0.006), and liver to 272±13; P = 0.08). Rates of readmission to the
failure as defined by alanine aminotransferase lev- ICU were similar (6.3 percent) in the two groups.
els above 250 U per liter (odds ratio, 1.62; 95 per-
cent confidence interval, 1.01 to 2.60; P = 0.04). Stays in ICU Longer Than Three Days
Among the 767 patients who stayed for more than
Morbidity three days in the ICU, there was no significant
Intention-to-Treat Population difference between the two groups in the use of
In the intention-to-treat population, there was no any medications other than insulin. Among the
significant difference between the two treatment 386 patients in the intensive-treatment group, in-
groups in the use of medications other than in- tensive insulin therapy for at least a third day, as
sulin. Of 1200 patients in the intention-to-treat compared with conventional therapy, accelerated
population, 9 were treated for septic shock with weaning from mechanical ventilation (hazard ra-
activated protein C, 5 in the conventional-treat- tio, 1.43; 95 percent confidence interval, 1.16 to
ment group and 4 in the intensive-treatment 1.75; P<0.001), discharge from the ICU (hazard
group (P = 0.8). Of 644 patients receiving cortico- ratio, 1.34; 95 percent confidence interval, 1.12
steroid therapy, 327 were in the conventional- to 1.61; P = 0.002), and discharge from the hospi-
treatment group and 317 were in the intensive- tal (hazard ratio, 1.58; 95 percent confidence in-
treatment group (P = 0.8). The corticosteroid therapy terval, 1.28 to 1.95; P<0.001) (Fig. 3).
consisted largely of immunosuppressive or anti- In the conventional-treatment group, 28.6 per-
inflammatory treatment with methylprednisolone cent of patients received dialysis therapy, as com-
(at a median dose of 40 mg [interquartile range, pared with 27.2 percent of those in the intensive-
24 to 75] per treatment day among 233 patients in treatment group (P = 0.7). The use of dialysis in
the conventional-treatment group and a median patients who did not require dialysis before ad-
dose of 40 mg [interquartile range, 29 to 65] per mission to the ICU was not significantly reduced
day among 249 patients in the intensive-treatment (22.7 percent in the conventional-treatment group
group; P>0.9). Hydrocortisone was given for pre- and 20.8 percent in the intensive-treatment group,
sumed adrenal failure at a median dose of 125 mg P = 0.5). However, acquired kidney injury occur-
(interquartile range, 100 to 193) per day to 129 ring after randomization, as defined by a serum
patients in the conventional-treatment group and creatinine level at least twice that present on ad-
at a median dose of 135 mg (interquartile range, mission to the ICU (12.6 percent in the conven-
100 to 240) per day to 118 patients in the inten- tional-treatment group and 8.3 percent in the in-
sive-treatment group (P = 0.2). Five patients, two tensive-treatment group, P = 0.05) and the fraction
in the conventional-treatment group and three in of patients reaching a peak serum creatinine level
the intensive-treatment group, received a median greater than 2.5 mg per deciliter (39.4 and 32.5
daily dose of 10 mg of dexamethasone (P>0.9). percent, respectively; P = 0.04), was reduced. Hyper-
Morbidity was reduced in the intensive-treat- bilirubinemia was present in 55.2 percent of pa-
ment group, as reflected by a reduction in newly tients in the conventional-treatment group and
acquired kidney injury (8.9 to 5.9 percent, P = 0.04) 47.3 percent of those in the intensive-treatment
and in earlier weaning from mechanical ventila- group (P = 0.04). The levels of alanine aminotrans-
tion, as compared with the conventional-treat- ferase or aspartate aminotransferase were similar
ment group (hazard ratio, 1.21; 95 percent con- in the two groups.

n engl j med 354;5 www.nejm.org february 2, 2006 455

The New England Journal of Medicine

Downloaded from nejm.org on April 30, 2014. For personal use only. No other uses without permission.
Copyright © 2006 Massachusetts Medical Society. All rights reserved.
 The new england journal of medicine

A
Weaning from Mechanical
Ventilation Discharge from ICU Discharge from Hospital
4.0 P=0.03 4.5 P=0.04 5.0 P=0.05
Cumulative Hazard

3.5 4.0
3.0 Intensive 3.5 4.0
2.5 treatment 3.0
2.5 3.0
2.0 2.0
1.5 2.0
Conventional 1.5
1.0 1.0 1.0
0.5 treatment
0.5
0.0 0.0 0.0
0 10 20 30 40 50 60 70 80 90 0 20 40 60 80 100 0 100 200 300 400 500 600
Days after Admission to ICU

B
Weaning from Mechanical
Ventilation Discharge from ICU Discharge from Hospital
3.5 P<0.001 4.0 P=0.002 5.0 P<0.001
Cumulative Hazard

3.0 3.5
Intensive 3.0 4.0
2.5
treatment 2.5 3.0
2.0
2.0
1.5 1.5 2.0
1.0 Conventional 1.0 1.0
0.5 treatment 0.5
0.0 0.0 0.0
0 10 20 30 40 50 60 70 80 90 0 20 40 60 80 100 0 100 200 300 400 500 600
Days after Admission to ICU

Figure 3. Effect of Intensive Insulin Therapy on Morbidity.

The effect of intensive insulin therapy on time to weaning from mechanical ventilation, time to discharge from the intensive care unit
(ICU), and time to discharge from the hospital is shown for all patients (intention-to-treat analysis, Panel A) and for the subgroup of 767
patients staying in the ICU for three or more days (Panel B). P values for the comparison between the two groups were calculated by
proportional-hazards regression analysis with censoring for early deaths. Circles represent patients.

The proportion of patients who had bacteremia also beneficially affected, with no effect among
(11.3 percent) or secondary bacteremia (7.3 per- those treated for less than five days.
cent) or received prolonged antibiotic therapy (37.6
percent in the conventional-treatment group and Mortality
31.9 percent in the intensive-treatment group,
P = 0.09) was not significantly reduced. However, Among the 1200 patients included in the inten-
intensive insulin therapy reduced the incidence of tion-to-treat analysis, ICU and in-hospital mor-
hyperinflammation from 74 percent in the conven- tality were not significantly reduced by intensive
tional-treatment group to 67 percent in the inten- insulin therapy (Table 2 and Fig. 4). For all patients,
sive-treatment group (P = 0.03). mortality in the ICU at day 3 (2.8 percent vs. 3.9
Intensive insulin therapy reduced the cumula- percent, P = 0.31) and in-hospital mortality at day
tive TISS-28 scores among patients in the ICU by 3 (3.6 percent vs. 4.0 percent, P = 0.72) were not sig-
20 percent (454±22 in the conventional-treatment nificantly different in the two treatment groups.
group vs. 388±17 in the intensive-treatment group, Beyond the third day of intensive insulin therapy,
P = 0.02), reflecting a reduction in the costs of in- the in-hospital mortality was reduced from 52.5
tensive care.19,20 Among patients who underwent to 43.0 percent (Fig. 4 and Table 2). Death from
randomization and stayed in the ICU less than all causes in the ICU appeared to be reduced. The
three days, none of the morbidity end points effect on mortality among patients staying for
were significantly different in the two treatment more than three days in the ICU was shown in
groups. Beyond the fifth day of intensive insulin most of the subgroups stratified according to di-
therapy, all the morbidity end points studied were agnostic category, but it was much less pronounced

456 n engl j med 354;5 www.nejm.org february 2, 2006

The New England Journal of Medicine

Downloaded from nejm.org on April 30, 2014. For personal use only. No other uses without permission.
Copyright © 2006 Massachusetts Medical Society. All rights reserved.
 Intensive Insulin Ther apy in the Medical ICU

Table 2. Mortality in the Study Groups.*

Variable Intention-to-Treat Group Group in ICU for ≥3 Days

Conventional Intensive P Value Conventional Intensive P Value
Treatment Treatment Treatment Treatment
(N = 605) (N = 595) (N = 381) (N = 386)
Total deaths during intensive care 162 (26.8) 144 (24.2) 0.31 145 (38.1) 121 (31.3) 0.05
— no. (%)
Causes of death during intensive care — 0.90 0.70
no. (% of patients in the category)
Persistent MOF after septic or 59 (51.3) 56 (48.7) 55 (51.4) 52 (48.6)
SIRS-induced shock
Respiratory failure 50 (54.3) 42 (45.7) 49 (55.7) 39 (44.3)
Therapy-resistant septic shock 21 (50.0) 21 (50.0) 14 (53.8) 12 (46.2)
Cardiovascular collapse 18 (54.5) 15 (45.5) 16 (66.7) 8 (33.3)
Severe brain damage 14 (58.3) 10 (41.7) 11 (52.4) 10 (47.6)
In-hospital deaths — no. (%) 242 (40.0) 222 (37.3) 0.33 200 (52.5) 166 (43.0) 0.009
Hazard ratio (95% CI) 0.94 (0.84–1.06) 0.31 0.84 (0.73–0.97) 0.02†
In-hospital deaths, according to diag-
nostic category — no. (% of
patients in the category)
Respiratory 115 (44.1) 98 (39.0) 100 (58.1) 78 (41.7)
Gastrointestinal or liver disease 49 (32.2) 41 (45.6) 39 (43.8) 23 (32.4)
Hematologic or oncologic disease 33 (64.7) 28 (60.9) 26 (70.3) 23 (58.9)
Other sepsis 13 (28.9) 19 (42.2) 11 (36.7) 16 (47.0)
Cardiovascular 8 (33.3) 14 (53.8) 6 (40.0) 11 (61.1)
Neurologic 9 (29.0) 9 (30.0) 6 (42.8) 5 (35.7)
Renal 8 (40.0) 6 (25.0) 7 (63.6) 4 (30.8)
Metabolic 4 (36.4) 2 (22.2) 4 (40.0) 2 (40.0)
Other 3 (30.0) 5 (50.0) 1 (33.3) 4 (80.0)
In-hospital deaths, according to
APACHE II quartile — no.
(% of patients in the category)
<17 25 (19.5) 28 (19.7) 25 (34.7) 20 (24.1)
17 to 22 63 (37.1) 54 (32.1) 51 (51.5) 38 (38.8)
23 to 29 74 (44.6) 66 (44.0) 63 (55.8) 51 (47.7)
>29 79 (58.1) 73 (55.3) 61 (62.9) 57 (58.2)
In-hospital deaths, according to history
of diabetes — no. (% of pa-
tients in the category)
No history of diabetes 208 (40.9) 180 (36.8) 173 (53.4) 137 (41.9)
History of diabetes 34 (35.0) 42 (39.6) 27 (47.4) 29 (49.2)
28-Day mortality — no. (%) 182 (30.0) 178 (29.9) 0.95 149 (39.1) 133 (34.5) 0.18
90-Day mortality — no. (%) 228 (37.7) 214 (35.9) 0.53 187 (49.1) 163 (42.2) 0.06
Deaths in ICU on day 3 — no. (%) 17 (2.8) 23 (3.9) 0.31
Deaths in hospital on day 3 — no. (%) 22 (3.6) 24 (4.0) 0.72

* ICU denotes intensive care unit, MOF multiple organ failure, SIRS systemic inflammatory response syndrome, and CI confidence interval.
P values were calculated by the chi-square test, uncorrected for the crude mortality data and corrected for baseline risk factors for the odds
ratios for in-hospital death obtained by proportional-hazards regression analysis. Clinically relevant baseline risk factors for death included
severity of illness scores (APACHE II and Therapeutic Intervention Scoring System-28 scores), a history of diabetes, active cancer and kidney
failure before admission to the ICU, dialysis dependence, signs of liver necrosis (alanine aminotransferase level, >150 IU per liter), baseline
plasma urea level >150 mg per deciliter, and hyperinflammation (C-reactive protein level, >150 mg per deciliter).
† The P value has been corrected for the risk factors.

n engl j med 354;5 www.nejm.org february 2, 2006 457

The New England Journal of Medicine

Downloaded from nejm.org on April 30, 2014. For personal use only. No other uses without permission.
Copyright © 2006 Massachusetts Medical Society. All rights reserved.
 The new england journal of medicine

A Intention-to-Treat Group (N=1200) B Subgroup in ICU ≥3 Days (N=767)

100 100

80 80

Intensive treatment
In-Hospital Survival (%)

In-Hospital Survival (%)

60 60 Intensive treatment
Conventional treatment

100 100
Conventional treatment
40 40
80 80

60 60
20 20

40 40
First 30 days First 30 days
0 0
0 10 20 30 0 10 20 30
0 0
0 100 200 300 400 500 0 50 100 150 200 250 300 350 500
Days Days

Figure 4. Kaplan–Meier Curves for In-Hospital Survival.

The effect of intensive insulin treatment on the time from admission to the intensive care unit (ICU) until death is
shown for the intention-to-treat group (Panel A) and the subgroup of patients staying in the ICU for three or more
days (Panel B). Patients discharged alive from the hospital were considered survivors. P values calculated by the
log-rank test were 0.40 for the intention-to-treat group and 0.02 for the subgroup staying in the ICU for three or
more days. P values calculated by proportional-hazards regression analysis were 0.30 and 0.02, respectively.

in the highest APACHE II quartile (Table 2). Among whom intensive care had been limited or with-
the 433 patients who stayed in the ICU less than drawn within 72 hours after admission to the ICU
three days and for whom data were censored af- (26 patients in the conventional-treatment group
ter randomization, 56 of those in the intensive- and 39 in the intensive-treatment group). Of these
treatment group and 42 in the conventional-treat- 65 patients, 29 had long stays in the ICU (16 in
ment group died, but the statistical significance the conventional-treatment group and 13 in the in-
of this finding varied depending on the analyti- tensive-treatment group), and 36 had short stays
cal approach (P = 0.05 with the chi-square test; (10 and 26, respectively). After censoring, the
hazard ratio, 1.09; 95 percent confidence interval, in-hospital mortality in the intention-to-treat pop-
0.90 to 1.32; P = 0.35 by uncorrected proportional- ulation was 37.8 percent in the conventional-treat-
hazards analysis; hazard ratio, 1.09; 95 percent ment group versus 33.5 percent in the intensive-
confidence interval, 0.89 to 1.32; P = 0.41 after cor- treatment group (P = 0.1); among those with long
recting for baseline risk factors listed in Table 2). stays in the ICU, the in-hospital mortality was 50.9
Beyond the fifth day of intensive insulin ther- percent versus 41.5 percent (P = 0.01); and among
apy, mortality was reduced from 54.9 to 45.9 per- those with short stays, it was 15.4 percent versus
cent (P = 0.03), with no significant effect among 16.9 percent (P = 0.7).
patients staying less than five days in the ICU
(P = 0.50). Dis cus sion

Post Hoc Exploratory Mortality Analysis Intensive insulin therapy during intensive care
Of the 1200 patients in the total study group, a prevented morbidity but did not significantly re-
post hoc analysis censored data on 65 patients for duce the risk of death among all patients in the

458 n engl j med 354;5 www.nejm.org february 2, 2006

The New England Journal of Medicine

Downloaded from nejm.org on April 30, 2014. For personal use only. No other uses without permission.
Copyright © 2006 Massachusetts Medical Society. All rights reserved.
 Intensive Insulin Ther apy in the Medical ICU

medical ICU included in the intention-to-treat compared with patients staying in the ICU for at
population. However, among those who stayed least three days, is that the benefit from inten-
in the ICU for three or more days, intensive insu- sive insulin therapy requires time to be realized.
lin therapy reduced morbidity and mortality. Indeed, the intervention is aimed not at curing
The reduced morbidity resulted from the pre- disease but at preventing complications that occur
vention of acquired kidney injury, earlier weaning during and, perhaps in part as a result of, inten-
from mechanical ventilation, and earlier discharge sive care. Prevention probably does not occur when
from the medical ICU and the hospital in patients the patient has a high risk of death from the dis-
who received intensive insulin therapy as com- ease causing admission to the ICU and when the
pared with those who did not. In contrast to pa- intervention is administered for a relatively short
tients in the surgical ICU,5 however, those in the time. However, among patients in whom compli-
medical ICU had no detectable reduction in bac- cations resulting from intensive care contribute
teremia, which may be explained by the fact that to an adverse outcome, such a preventive strategy,
among medical patients sepsis often triggers ad- if continued long enough, is likely to be effective.
mission to the ICU, irrespective of the disease ne- This would explain why patients with long stays in
cessitating hospital admission. Although infec- the medical ICU benefit more than those with
tions other than bacteremia were not analyzed for short stays, as shown in a surgical ICU.5
our study and may have been missed, the anti- Among patients staying for at least three days
inflammatory effect8 and the protection of organ in the ICU, the absolute reduction in in-hospital
function9 appeared to be independent of preven- mortality associated with intensive insulin ther-
tion of infection. Possible mechanisms of action apy was similar to that in our previous report5
include the prevention of cellular hypoxia by means and exceeded the effect on mortality in the ICU,
of reduced endothelial damage10 and the preven- indicating that intensive insulin therapy during
tion of cytopathic hypoxia.11 intensive care had a carryover effect. Such a lon-
Analysis of the subgroup treated in the ICU ger-term effect is in line with our previous find-
for three or more days showed not only a benefi- ing of superior long-term rehabilitation among
cial effect on morbidity but also a reduction in patients with brain injury who received intensive
mortality that was absent in the total study popu- insulin therapy during intensive care.24 Mortality
lation. However, since the length of stay in the ICU in a subgroup with a diagnosis of diabetes ap-
cannot be predicted for an individual patient and peared to be unaffected by intensive insulin ther-
therefore the analysis based on length of stay in- apy, although the numbers were small. This find-
evitably requires post-randomization stratifica- ing may be explained in part by the fact that the
tion, there is a risk of bias. It is unclear whether target blood glucose level was not reached in this
intensive insulin therapy received for less than subgroup. Indeed, achieving normoglycemia ap-
three days caused harm, as might be inferred from pears crucial to obtaining the benefit of intensive
the greater number of deaths among patients insulin therapy.6
staying less than three days in the ICU. Post hoc In the present study, normoglycemia was
exploratory analysis, with its inherent limitations, achieved with insulin titrated by the attending
suggested that this apparent difference in mor- nurses in the ICU. Despite the use of guidelines
tality among those staying a shorter time in the similar to those used in the surgical study,5 an epi-
ICU could be explained by the higher number of sode of biochemical hypoglycemia occurred more
patients in the intensive-treatment group for whom often among the patients in the medical ICU. Liver
intensive care was limited or withdrawn for rea- failure and kidney failure, which increase the vul-
sons of futility within 72 hours after admission. nerability to hypoglycemia, may partly explain this
In our previous study, brief exposure to insulin observation. However, logistic-regression analysis
therapy had no significant effect on the risk of identified hypoglycemia as an independent risk
death.5 Why 48 hours or less of insulin therapy factor for death. Hence, it is possible that hypogly-
would cause harm, whereas sustained treatment cemia induced by intensive insulin therapy may
would be beneficial, is unclear. have reduced a portion of the potential benefit.
An alternative and more likely explanation for This study has certain limitations. Like our
the difference in the effect of intensive insulin previous study of the surgical ICU,5 this was a
therapy in the intention-to-treat population, as single-center study; and as in previous studies in

n engl j med 354;5 www.nejm.org february 2, 2006 459

The New England Journal of Medicine

Downloaded from nejm.org on April 30, 2014. For personal use only. No other uses without permission.
Copyright © 2006 Massachusetts Medical Society. All rights reserved.
 The new england journal of medicine

patients with diabetes mellitus,25,26 it was not pos- Thus, targeting blood glucose levels to below
sible to achieve strict blinding, because safe insu- 110 mg per deciliter with insulin therapy prevented
lin titration requires monitoring of blood glucose morbidity but did not significantly reduce mor-
levels. However, because physicians on the gen- tality among all patients in our medical ICU. How-
eral wards were unaware of the treatment as- ever, intensive insulin therapy in patients who
signments of patients receiving intensive care and stayed in the ICU for at least three days was as-
had no access to the results of blood glucose test- sociated with reduced morbidity and mortality.
ing, bias in the analysis of the effect on length Large multicenter trials are needed to confirm
of stay in the hospital and in the analysis of in- these preliminary results.
hospital mortality was prevented. Furthermore, Supported by grants from the Belgian Fund for Scientific Re-
search (G.0278.03 and G.3C05.95N), the Research Council of the
since there was no survival benefit in the inten- University of Leuven (OT/03/56), and the Belgian Foundation for
tion-to-treat group, as compared with the sub- Research in Congenital Heart Diseases (to Dr. Van den Berghe).
group staying in the ICU for three or more days, Dr. Van den Berghe reports having received an unrestricted
research grant from Novo Nordisk; and Dr. Bouillon, an unre-
the use of intensive insulin therapy in all patients stricted research grant from Servier. No other potential conflict
in the medical ICU, including those staying less of interest relevant to this article was reported.
than three days, could be questioned. Because pa- We are indebted to Mona Eerdekens, M.Sc., and Peggy Claes,
R.N., for assistance with blood samples and data collection; to
tients who will have a prolonged stay in the ICU the clinical fellows in the ICU for APACHE II scoring; to the
cannot be identified with certainty on admission, nurses for daily TISS-28 scoring and for excellent compliance
adequately powered trials are needed to address with the protocol; to Prof. Emmanuel Lesaffre for inspiring dis-
cussions of the statistical analysis; and to HemoCue, Ängel-
this important issue. On the basis of our current holm, Sweden, for generously providing the equipment and re-
data, such studies would require at least 5000 pa- agents for point-of-care blood glucose measurements.
tients in the medical ICU.

References
1. Mizock BA. Alterations in carbohy- nose-binding lectin levels. J Clin Endocri- 2): effects on mortality and morbidity.
drate metabolism during stress: a review of nol Metab 2003;88:1082-8. Eur Heart J 2005;26:650-61.
the literature. Am J Med 1995;98:75-84. 9. Van den Berghe G. How does blood 16. Furnary AP, Gao G, Grunkemeier GL,
2. McCowen KC, Malhotra A, Bistrian BR. glucose control with insulin save lives in et al. Continuous insulin infusion reduces
Stress-induced hyperglycemia. Crit Care intensive care? J Clin Invest 2004;114:1187- mortality in patients with diabetes under-
Clin 2001;17:107-24. 95. going coronary artery bypass grafting.
3. Capes SE, Hunt D, Malmberg K, Ger- 10. Langouche L, Vanhorebeek I, Vlasse- J Thorac Cardiovasc Surg 2003;125:1007-21.
stein HC. Stress hyperglycaemia and in- laers D, et al. Intensive insulin therapy 17. Souba WW. Nutritional support.
creased risk of death after myocardial in- protects the endothelium of critically ill N Engl J Med 1997;336:41-8.
farction in patients with and without patients. J Clin Invest 2005;115:2277-86. 18. Knaus WA, Draper EA, Wagner DP,
diabetes: a systematic overview. Lancet 11. Vanhorebeek I, De Vos R, Mesotten Zimmerman JE. APACHE II: a severity of
2000;355:773-8. M, Wouters PJ, De Wolf-Peeters C, Van disease classification system. Crit Care
4. Cely CM, Arora P, Quartin AA, Kett den Berghe G. Protection of hepatocyte Med 1985;13:818-29.
DH, Schein RMH. Relationship of base- mitochondrial ultrastructure and function 19. Miranda DR, de Rijk A, Schaufeli W.
line glucose homeostasis to hyperglyce- by strict blood glucose control with insu- Simplified Therapeutic Intervention Scor-
mia during medical critical illness. Chest lin in critically ill patients. Lancet 2005; ing System: the TISS-28 items — results
2004;126:879-87. 365:53-9. from a multicenter study. Crit Care Med
5. Van den Berghe G, Wouters P, Week- 12. Evans TW. Hemodynamic and meta- 1996;24:64-73.
ers F, et al. Intensive insulin therapy in bolic therapy in critically ill patients. 20. Keene AR, Cullen DJ. Therapeutic In-
critically ill patients. N Engl J Med 2001; N Engl J Med 2001;345:1417-8. tervention Scoring System: update 1983.
345:1359-67. 13. Clement S, Braithwaite SS, Magee MF, Crit Care Med 1983;11:1-3.
6. Van den Berghe G, Wouters PJ, Bouil- et al. Management of diabetes and hyper- 21. Weinstein MP, Towns ML, Quartey
lon R, et al. Outcome benefit of intensive glycemia in hospitals. Diabetes Care 2004; SM, et al. The clinical significance of posi-
insulin therapy in the critically ill: insulin 27:553-91. [Errata, Diabetes Care 2004;27: tive blood cultures in the 1990s: a pro-
dose versus glycemic control. Crit Care 856, 2004;27:1255.] spective comprehensive evaluation of the
Med 2003;31:359-66. 14. Malmberg K, Ryden L, Efendic S, et microbiology, epidemiology, and outcome
7. Weekers F, Giulietti A-P, Michalaki M, al. Randomized trial of insulin-glucose of bacteremia and fungemia in adults. Clin
et al. Metabolic, endocrine, and immune infusion followed by subcutaneous insulin Infect Dis 1997;24:584-602.
effects of stress hyperglycemia in a rabbit treatment in diabetic patients with acute 22. Weinstein MP, Mirrett S, Van Pelt L, et
model of prolonged critical illness. Endo- myocardial infarction (DIGAMI study): ef- al. Clinical importance of identifying co-
crinology 2003;144:5329-38. fects on mortality at 1 year. J Am Coll Car- agulase-negative staphylococci isolated
8. Hansen TK, Thiel S, Wouters PJ, diol 1995;26:57-65. from blood cultures: evaluation of MicroScan
Christiansen JS, Van den Berghe G. Inten- 15. Malmberg K, Ryden L, Wedel H, et al. Rapid and Dried Overnight Gram-Positive
sive insulin therapy exerts anti-inflam- Intense metabolic control by means of in- panels versus a conventional reference
matory effects in critically ill patients and sulin in patients with diabetes mellitus method. J Clin Microbiol 1998;36:2089-92.
counteracts the adverse effect of low man- and acute myocardial infarction (DIGAMI 23. Bellomo R, Ronco C, Kellum JA, et al.

460 n engl j med 354;5 www.nejm.org february 2, 2006

The New England Journal of Medicine

Downloaded from nejm.org on April 30, 2014. For personal use only. No other uses without permission.
Copyright © 2006 Massachusetts Medical Society. All rights reserved.
 Intensive Insulin Ther apy in the Medical ICU

Acute renal failure: definition, outcome therapy protects the central and peripheral 26. UK Prospective Diabetes Study Group.
measures, animal models, fluid therapy nervous system of intensive care patients. Intensive blood-glucose control with sul-
and information technology needs: the Neurology 2005;64:1348-53. phonylureas or insulin compared with
Second International Consensus Confer- 25. Diabetes Control and Complications conventional treatment and risk of com-
ence of the Acute Dialysis Quality Initia- Trial Research Group. The effect of inten- plications in patients with type 2 diabetes
tive (ADQI) Group. Crit Care 2004;8: sive treatment of diabetes on the develop- (UKPDS 33). Lancet 1998;352:837-53. [Er-
R204–R212. ment and progression of long-term com- ratum, Lancet 1999;354:602.]
24. Van den Berghe G, Schoonheydt K, plications in insulin-dependent diabetes Copyright © 2006 Massachusetts Medical Society.
Becx P, Bruyninckx F, Wouters PJ. Insulin mellitus. N Engl J Med 1993;329:977-86.

n engl j med 354;5 www.nejm.org february 2, 2006 461

The New England Journal of Medicine

Downloaded from nejm.org on April 30, 2014. For personal use only. No other uses without permission.
Copyright © 2006 Massachusetts Medical Society. All rights reserved.