New England Journal of Medicine MARCH 26, 2009

Intensive versus Conventional Glucose Control in Critically Ill Patients

The NICE-SUGAR Study Investigators (Simon Finfer, et al.)

Bottom Line

In this large, international, randomized controlled trial, intensive glucose control (target 81–108 mg/dL) in critically ill adults resulted in increased 90-day mortality and a significantly higher incidence of severe hypoglycemia compared to conventional glucose control (target ≤180 mg/dL).

Key Findings

1. Intensive glucose control increased 90-day mortality compared to conventional control (27.5% vs. 24.9%; odds ratio 1.14, 95% CI 1.02 to 1.28, P=0.02).

2. The absolute risk of death was 2.6 percentage points higher in the intensive group, corresponding to a number needed to harm (NNH) of 38.

3. Severe hypoglycemia (blood glucose ≤40 mg/dL) was significantly more frequent in the intensive control group compared to the conventional group (6.8% vs. 0.5%; P<0.001).

4. There were no significant differences between the study groups in the duration of mechanical ventilation, length of stay in the ICU or hospital, or the need for renal replacement therapy.

Study Design

Design

RCT

Open-Label

Sample

6,104

Patients

Duration

90 days

Median

Setting

Multicenter, international

Population Adult patients admitted to the ICU expected to require intensive care for 3 or more consecutive days.

Intervention Intensive glucose control targeting blood glucose between 81 and 108 mg/dL.

Comparator Conventional glucose control targeting blood glucose of 180 mg/dL or less.

Outcome Death from any cause within 90 days after randomization.

Study Limitations

• The open-label nature of the trial, where clinicians were aware of group assignments, may have introduced bias.

• There was an imbalance in the use of corticosteroids between the groups, which might have influenced mortality outcomes.

• The intensive target range of 81–108 mg/dL was difficult to maintain consistently, as evidenced by the higher median blood glucose levels achieved in the intensive arm compared to the protocol target.

• Premature discontinuation of study treatment occurred more frequently in the intensive group, potentially introducing selection or attrition bias.

Clinical Significance

The trial challenged the prevailing belief that 'tight' glycemic control in the ICU was beneficial. By demonstrating increased mortality and morbidity, it provided strong evidence against aggressive glucose targets, leading to a shift in clinical practice toward more conservative glycemic management strategies (target ≤180 mg/dL) in critically ill patients.

Historical Context

The study was prompted by conflicting data, most notably the 2001 Leuven I trial, which had suggested that intensive insulin therapy significantly reduced morbidity and mortality in surgical ICU patients. The uncertainty surrounding those findings led the critical care community to design the robust, multicenter NICE-SUGAR trial to definitively test the safety and efficacy of intensive glycemic control.

Guided Discussion

High-yield insights from every perspective

Med Student

Medical Student

Why does 'stress hyperglycemia' occur in critically ill patients without a history of diabetes, and based on NICE-SUGAR, why is it dangerous to aggressively correct it to 'normal' levels (80-110 mg/dL)?

Key Response

Stress-induced hyperglycemia is a physiologic response to counter-regulatory hormones (cortisol, catecholamines) and cytokines that increase gluconeogenesis and insulin resistance. NICE-SUGAR demonstrated that while hyperglycemia is associated with poor outcomes, aggressively normalizing it significantly increases the risk of severe hypoglycemia (glucose <40 mg/dL), which is neuroglycopenic and associated with increased mortality.

Resident

In a patient with septic shock and blood glucose readings consistently between 150 and 170 mg/dL, how do the findings of the NICE-SUGAR trial guide your decision to initiate an insulin drip?

Key Response

The NICE-SUGAR trial compared intensive (81-108 mg/dL) to conventional control (target <180 mg/dL). It showed that the conventional approach resulted in lower 90-day mortality. Therefore, in a patient stable at 150-170 mg/dL, insulin therapy is not indicated as it provides no survival benefit and increases the risk of life-threatening hypoglycemia.

Fellow

How did the discrepancy between the 2001 Leuven trial and the NICE-SUGAR trial regarding the surgical ICU population change our understanding of glucose management in the perioperative setting?

Key Response

The initial Leuven trial suggested a mortality benefit for intensive control in surgical patients, but it was a single-center study that heavily utilized parenteral nutrition. NICE-SUGAR, a much larger multicenter trial, found that the harm of intensive control (higher mortality) persisted across both medical and surgical subgroups, leading to a global shift away from tight glycemic targets even in postoperative care.

Attending

NICE-SUGAR is often cited as a landmark trial in 'de-implementation' science. How does this study challenge the use of surrogate physiological endpoints in the design of ICU protocols?

Key Response

For years, clinicians assumed that normalizing glucose (a physiological surrogate) would improve outcomes based on observational data linking hyperglycemia to death. NICE-SUGAR proved that achieving a 'normal' lab value via intensive intervention can be more harmful than the underlying pathology. It teaches that ICU interventions must be validated by patient-centered outcomes (mortality) rather than just physiologic normalization.

Scholarly Review

Critical appraisal through the lens of expert reviewers and guideline development

PhD

In the statistical analysis of NICE-SUGAR, how should we approach the potential for 'hypoglycemia-associated mortality' to be a marker of illness severity rather than a direct causal mechanism for death?

Key Response

This addresses the issue of residual confounding. While hypoglycemia correlates with death, the most critically ill patients (higher APACHE scores) often have the most glucose lability. A PhD-level critique would look for time-dependent covariate analysis or mediation analysis to determine if hypoglycemia independently mediates the relationship between treatment arm and death, or if it is simply a reflection of the patient's underlying physiological fragility.

Journal Editor

If you were reviewing the NICE-SUGAR manuscript, how would you evaluate the 'conventional' arm's achieved mean glucose of ~145 mg/dL when their target was <180 mg/dL, and what implications does this 'mid-range' control have on the trial's generalizability?

Key Response

A critical reviewer would note that the 'conventional' arm was actually quite well-controlled (not neglected). This 'mid-range' achievement suggests that the study didn't compare 'good vs. bad' control, but rather 'intensive vs. moderate' control. This narrows the separation between groups but reinforces the safety and efficacy of avoiding the 80-110 mg/dL range in a real-world pragmatic setting.

Guideline Committee

How did the NICE-SUGAR results influence the Surviving Sepsis Campaign's recommendations for glucose triggers and targets, and what is the current strength of evidence for the 180 mg/dL threshold?

Key Response

NICE-SUGAR is the primary evidence for the Surviving Sepsis Campaign's Grade 1A recommendation to initiate insulin therapy only when blood glucose levels are >180 mg/dL, with a target of 144–180 mg/dL. This represented a direct reversal of previous guidelines that had leaned toward tighter control based on smaller, lower-quality studies.

Clinical Landscape

Noteworthy Related Trials

2001

Leuven I Study

n = 1,548 · NEJM

Tested

Intensive insulin therapy (target 80-110 mg/dL)

Population

Critically ill patients in a surgical ICU

Comparator

Conventional insulin therapy (target 180-200 mg/dL)

Endpoint

In-hospital mortality

Key result: Intensive insulin therapy reduced mortality and morbidity in surgical ICU patients.

2008

ACCORD Trial

n = 10,251 · NEJM

Tested

Intensive glucose control (HbA1c < 6.0%)

Population

Patients with type 2 diabetes and high cardiovascular risk

Comparator

Standard glucose control (HbA1c 7.0-7.9%)

Endpoint

Composite of nonfatal MI, nonfatal stroke, or cardiovascular death

Key result: Intensive therapy increased mortality and did not significantly reduce major cardiovascular events.

2009

NICE-SUGAR Study

n = 6,039 · NEJM

Tested

Intensive glucose control (target 81-108 mg/dL)

Population

Critically ill adults in the ICU

Comparator

Conventional glucose control (target <= 180 mg/dL)

Endpoint

90-day all-cause mortality

Key result: Intensive glucose control increased mortality compared to conventional control in critically ill patients.

Tailored to your role

Want this tailored to you?

Add your specialty or training stage to get role-specific takeaways and more questions.

Personalize this analysis