Hyperglycemia management in the ICU

“Intensive versus conventional glucose control in critically ill patients.” – the NICE-SUGAR investigators, The New England Journal of Medicine, March 26th, 2009

“Intensive insulin therapy in critically ill patients.” – Greet van de Berghe et al., The New England Journal of Medicine, November 8, 2001

“Intensive insulin therapy in the medical ICU.” – Greet van de Berghe et al., The New England Journal of Medicine, February 2, 2006


Control of blood glucose in the Intensive Care Unit (ICU) and the hospital has implications in many disease processes, including cardiovascular, renal, and infectious problems. Elevated or abnormally low blood glucose values can compound with the primary problem and complicate a patient’s hospital stay. Over the course of the 2000s, three large studies attempted to establish and validate a strategy to control blood sugar in the ICU.

The first two trials (van de Berge 2001 and van de Berghe 2006) were conducted at a single center with patient numbers in the 1400-1500 range. The 2001 study followed patients in the Surgical ICU, and the 2006 study in the Medical ICU. These investigators proposed an intensive glucose control regimen where an insulin infusion was initiated at levels higher than 110 mg/dL (the upper limit of normal blood sugars) and titrated to blood levels in the 80-110 mg/dL normoglycemic range. This was compared with a conventionally treated group where insulin drips were started once blood sugar exceeded 215 mg/dL, and titrated to a range of 180-215 mg/dL.

The NICE-SUGAR study was a multi-centered study that included medical and surgical ICUs, with a total patient population of 6104. The intensive therapy was repeated similar to the studies above. The conventional arm of patients received an insulin drip once blood glucose levels exceeded 180 mg/dL. Insulin drip was discontinued when glucose levels fell below 144 mg/dL. The target glucose level was < 180 mg/dL. Importantly, inclusion criteria for this trial selected patients that expected to remain in the ICU > 3 days.


Van de Berghe 2001 found that the intensive therapy resulted in reduced in-hospital mortality and ICU mortality, especially in patients staying longer that 5 days. The intensive therapy group also had fewer morbidity rates including lower rates of sepsis. Hypoglycemia did occur more frequently in the intensive therapy group.

Van de Berghe 2006 also found ICU and in-hospital death was lower in the intensive treatment arm in patients who stayed in the ICU for longer than 3 days. There was no significance in mortality between the two arms in terms of in-hospital or ICU mortality for all ICU patients. Hypoglycemia occurred more frequently in the intensive therapy arm. There was an improvement in morbidities – requirements of mechanical ventilation, ICU stay and hospital stay in the intensive arm, but no significant fewer episodes of sepsis.

The NICE-SUGAR study found an increased all-cause mortality at 90 days after admission to the ICU in the intensive treatment arm when compared to conventional treatment. The majority of the deaths in both arms of the study were in-hospital or in the ICU. There was no significant difference in morbidities, including sepsis, except for an increased number of hypoglycemic episodes in the intensive glucose management group.

Why We Do What We Do

After the 2001 van de Berghe paper, intensive glucose management became the standard of practice in the ICU. However, after the striking results of the NICE-SUGAR study, the recommended practice is now a liberal approach to glucose management with a goal of blood sugars < 180 mg/dL and treatment only above this level. Large sample size and diversity in multiple trial centers provide this study with validity in a broad range of ICU and hospital applications.

Statistical significance of the data is also important – the 2006 medical ICU van de Berghe study failed to find a difference between in-hospital and ICU mortality between the two arms of the study for all patients, so the contemporary standard of practice (intensive management) was considered to be safe and valid. However, NICE-SUGAR’s results were statistically significant in showing that intensive therapy actually led to increased mortality. Clinical practice changed quickly as a result of this significant data.

After the results of the NICE-SUGAR study, there was extensive discussion into why intensive glucose control increased mortality, in stark contrast to the previous two landmark studies by van de Berghe. The authors of NICE-SUGAR did not expand on a cause for the increased mortality, but referred to lower blood sugars, increased insulin administration and increased episodes of hypoglycemia as being possible explanations. When under stress, as critical patients are, the body naturally produces a hyperglycemic state with increased corticosteroid responses, and dampening this response with artificial insulin administration may work against the complex defense mechanisms of the stressed-state body. The results of NICE-SUGAR interestingly correlate with another landmark trial on outpatient diabetes, the ACCORD trial, which also found that intensive glucose management increased mortality. However, more research was requested by both the ACCORD and NICE-SUGAR studies to explain their results.

1. NICE-SUGAR Study Investigators, Finfer S, Chittock DR, Su SY, Blair D, Foster
D, Dhingra V, Bellomo R, Cook D, Dodek P, Henderson WR, Hébert PC, Heritier S,
Heyland DK, McArthur C, McDonald E, Mitchell I, Myburgh JA, Norton R, Potter J,
Robinson BG, Ronco JJ. Intensive versus conventional glucose control in
critically ill patients. N Engl J Med. 2009 Mar 26;360(13):1283-97. doi:
10.1056/NEJMoa0810625. Epub 2009 Mar 24. PubMed PMID: 19318384.

2. van den Berghe G, Wouters P, Weekers F, Verwaest C, Bruyninckx F, Schetz M,
Vlasselaers D, Ferdinande P, Lauwers P, Bouillon R. Intensive insulin therapy in
critically ill patients. N Engl J Med. 2001 Nov 8;345(19):1359-67. PubMed PMID:

3. Van den Berghe G, Wilmer A, Hermans G, Meersseman W, Wouters PJ, Milants I,
Van Wijngaerden E, Bobbaers H, Bouillon R. Intensive insulin therapy in the
medical ICU. N Engl J Med. 2006 Feb 2;354(5):449-61. PubMed PMID: 16452557.


Stroke, t-PA and the NINDS trial

“Tissue Plasminogen Activator for Acute Ischemic Stroke” – The National Institute for Neurologic Disorders and Stroke rt-PA Stroke Study Group, New England Journal of Medicine, December 14th, 1995


Prior to the advent of intravenous thrombolytics, there was no direct treatment for ischemic strokes. The clot occluding the artery was allowed to sit in place and the area of brain deprived of blood flow, nutrients and oxygen was allowed to expand while the associated risk factors were controlled. Eventually the body would dissolve the clot (ironically, with it’s own “homemade” version of tissue plasminogen activator, or t-PA) and start the long road to recovery of brain function.

The difficulty with strokes is not mortality. In fact, people who die from large strokes due to massive clots are often dead before they reach a hospital or care facility. Preventing stroke mortality was only one component of the analysis conducted by these authors. Strokes cause disability. Depending on the region of the brain affected, strokes can cause weakness, inability to speak, inability to control emotions or any other mental function without killing, and often this is the most devastating result. Additionally, once disability has set in, the only treatment available is long term physical therapy, which is usually accompanied with large costs and uncertain outcomes. It is important to also note that there is a proportion of stroke patients who regain function of the affected brain region  without intervention. Often this indicates a smaller or less severe stroke, but even with advanced imaging at the initial assessment, it is difficult to characterize which patients will follow this course.


This study assessed the risks and benefits of administering intravenous t-PA within 90 and within 180 minutes of onset of stroke symptoms. The patients were assessed at 24 hours after stroke and 3 months after stroke. The tools for assessment were a variety of scales that measured disability. For example, the NIH Stroke Scale looks at multiple physical exam findings, such as strength in one extremity, to determine how much brain function the patient has lost. Mortality was also compared after 3 months.

The following results from the study were statistically significant: In comparing groups who received t-PA within 90 minutes and placebo, more t-PA patients than placebo patients showed significant improvement within the first 24 hours. After 3 months, more patients who received t-PA within 180 minutes were disability free than patients who received placebo, and this result was the same when patient groups were broken down by age and type and severity of stroke. There was an increased risk of symptomatic intracranial bleeding in patients given t-PA during the first 36 hours, but there was no difference in mortality between patients receiving t-PA and placebo after 3 months.

Why We Do What We Do

This study clearly showed that t-PA increases the chance of eliminating disability after three months if it is administered in the first 3 hours after symptom onset. The authors of this study chose a very difficult outcome measurement – the total reduction of disability to baseline, regardless of how severe the stroke was initially. By validating t-PA in this setting, they make a very strong case for its general use in ischemic strokes across the board.

Intravenous t-PA was the first direct treatment of stroke and is still, 18 years after the NINDS trial, the only effective therapy that we have. It has trounced intravascular mechanical clot-busting, numerous surgical procedures and many other drugs, therapies and delivery mechanisms. Most importantly, it works. For the patients who emerge with no disability, it is not just a therapy, it is a stroke cure. The NINDS study did not just validate the use and effectiveness of t-PA. For those who understand the metrics and analysis properly, it is a sign of hope that someday we will be able to eliminate, and not just reduce, the debilitating disability that results from stroke.

Tissue plasminogen activator for acute ischemic stroke. The National Institute
of Neurological Disorders and Stroke rt-PA Stroke Study Group. N Engl J Med. 1995
Dec 14;333(24):1581-7. PubMed PMID: 7477192.

A Brief Introduction

Welcome to Why We Do What We Do: A Guide to Landmark Studies in Medicine, a blog designed to showcase and discuss historically significant research studies that altered the landscape of medicine. Our goal is to create an arena, a virtual journal club of sorts, where medical students, MDs, healthcare professionals and even the casual reader can get a glimpse into the history behind the defining moments of medicine.

This is in no way a chronological compendium of clinical medicine; the posts are in no specific order, much like the clinical vignettes you may find published in medical journals. This is a means to pique your interest in the history behind our clinical algorithms and standard practices. In turn, I hope you gain a new appreciation for the MD-cum-novelist such as Siddhartha Mukherjee, Abraham Verghese or Allan Hamilton who have written at length on some of these same topics. Ultimately, we strive to familiarize you with the primary source for select, critical moments in medicine – those so blandly described in medical textbooks through which many of us have suffered. In our short clinical experience, curiosity got the best of us. Didn’t you ever wonder why we do what we do?