ICU reduces mortality of critically ill ICU patients

After initial resistance, benefits are great

When reading a New England Journal of Medicine article1 describing intensive monitoring of blood glucose levels for surgical intensive care unit (ICU) patients on ventilators, which resulted in decreased mortality rates, James Krinsley, MD, director of critical care at Stamford (CT) Hospital, had a revelation.

"When I read the article, the light bulb went on," he says. However, the study’s patient population differed from Krinsley’s ICU, as it included 63% of patients post-cardiac surgery.

"I have a 14-bed medical/surgical/cardiac ICU, and we don’t do cardiac surgery. So it’s a different population. But I wondered what the relationship was between mortality and glucose levels during admission to my ICU," he says.

This was the first question Krinsley asked of the ICU’s database, which he created in 1998 and now contains detailed information for more than 6,000 consecutive admissions. It turned out that there was a very strong correlation between mortality and glucose levels, with hospital mortality increasing progressively as glucose values increased, reaching 42.5% for patients with mean glucose values greater than 300 mg/dL.

"My nursing director heard me screaming through the wall that separates our offices; it was so exciting," he says.

Krinsley published his own findings detailing the strong relationship between increasing glucose levels during ICU admission and mortality in December 2003.2

Next, he set out to duplicate the study’s results in his own ICU. He began by talking with nurses about implementing a protocol of intensive monitoring and treatment to maintain normal blood glucose levels in all ICU patients. "I started a process of negotiation to find out what level the nurses were comfortable treating," Krinsley adds.

Initially, he proposed doing a randomized controlled study targeting glucose levels higher than 110. "Frankly, that crashed and burned because we don’t have sufficient resources for that type of study," Krinsley adds. "We don’t have research nurses or fellows; it’s basically me and the nurses, and there was also tremendous resistance to the low treatment goal."

After a great deal of discussion, the staff decided to target all glucose levels greater than 140. The new protocol was initiated in February 2003. "After two years of success with the original protocol, I am proud to say that in January 2005, we lowered our threshold from 140 to 125 — a change suggested by the nurses themselves," he reports. "Our mean glucose level, which was 152 in the baseline period and 129 during the original treatment period, is now down to 117, which is a phenomenal result."

In ICUs, the standard of care has been to ignore glucose levels as high as 175 to 225, and turning this around requires a whole new mindset — and additional work by nurses, Krinsley says. "No one treated elevated glucose because it was so common, and it was considered to be a normal and perhaps even beneficial stress response to acute, severe illness. It’s hard work — with extra monitoring, more decision making by the nurses, a degree of patient discomfort with all the finger sticks, and also the fear of hypoglycemia."

Krinsley continually produces charts and graphs to track his unit’s progress. "When we succeeded in meeting the new treatment threshold, colorful graphs were posted immediately in every nursing station in the ICU," he says.

The work paid off with dramatic results — mortality rates reduced by 29% and length of stay in the ICU reduced by nearly 11%, and also national recognition when the organization received the JCAHO’s 2004 Ernest A. Codman award.

"I’m incredibly proud of our nursing staff," adds Krinsley. "Fear of hypoglycemia has not materialized — there was no increase in glucose levels greater than 40 during the treatment period. In fact, during the first month of using the new protocol with 125 as our target, our glucose levels below 40 have actually gone down."3

The newest version of the protocol is simpler and gives nurses more leeway for clinical decision making. "We have created a cadre of excellent bedside glucose managers," he says. "They have flexibility within the parameters of the protocol to make individual treatment decisions, so we have empowered them medically."

Tight glucose control is a standard of care that many ICUs are struggling with, Krinsley acknowledges.

"It’s not easy to do. You need a protocol-driven culture, a data management system of some sort, and nursing and medicine on the same page," he explains. "We had resistance in the beginning, but we evolved to the point where we could conquer that and move forward."

Despite the difficulties, ICUs ultimately will adopt this practice universally, Krinsley predicts. "It’s safe and clinically effective. I am working on an analysis, which will prove that there are dramatic reductions in the cost of care because patients do better," he adds. "It isn’t one of the JCAHO core measures now, but could it become one? I can envision that."


1. Van den Berghe G, Wouters P, Weekers F, et al. Intensive insulin therapy in the critically ill patient. N Engl J Med 2001; 345:1,359-1,367.

2. Krinsley JS. Association between hyperglycemia and increased hospital mortality in a heterogeneous population of critically ill patients. Mayo Clin Proc 2003; 78:1,471-1,478.

3. Krinsley JS. Effect of an intensive glucose management protocol on the mortality of critically ill adult patients. Mayo Clin Proc 2004; 79:992-1,000.

[For more information, contact:

James Krinsley, MD, Director, Critical Care, Stamford Hospital, 30 Shelburne Road, Stamford, CT 06904-9317. Phone: (203) 348-2437. Fax: (203) 276-7243. E-mail:]