WASHINGTON — More hospitals are implementing standardized insulin infusion protocols, many of which emulate the Yale protocol, Dr. Philip A. Goldberg said at a consensus conference sponsored by the American Association of Clinical Endocrinologists, American College of Endocrinology, and the American Diabetes Association.
Dr. Goldberg, a postdoctoral fellow at Yale University, New Haven, Conn., said the protocol was introduced in 2001 after the publication of the landmark Leuven, Belgium study (N. Engl. J. Med. 2001;345:1359–67). Until then, the “state of the art” in the intensive care unit had been to tolerate blood glucose levels as long as they did not exceed 200 mg/dL and to rarely address plasma glucose elevations. Glucose levels were rarely checked in nondiabetic patients, and existing “sliding scale” insulin orders took into account only the current blood glucose. In contrast, the Yale protocol incorporated two other essential elements: The velocity of change (based on both the current and previous values) and the current infusion rate. “If you don't incorporate all three of those, your drip will not be successful,” he said.
In the first 69 insulin drips used in 52 medical ICU patients with a baseline mean glucose of 299 mg/dL, the median time to achieve target blood glucose levels of 100–139 mg/dL (now 90–119 mg/dL) was 9 hours, and the median drip duration was 61 hours. The protocol worked equally well in diabetic and nondiabetic patients, and was not influenced by the severity of illness.
The protocol was complex enough to achieve strict glycemic control in critically ill patients and practical enough to be implemented by busy ICU nurses without the need for continuous expert supervision (Diabetes Care 2004;27:461–7). Importantly, the protocol also was readily accepted by the nursing staff, with 73% rating it as “very easy” or “somewhat easy” to use.
“It's only complex the first two or three times you do it. Once you actually run an ICU nurse through this protocol a few times, it's not complex at all compared to the other things they're doing,” said Dr. Goldberg.
Since then, other institutions have created their own versions of the Yale protocol—some including computerized algorithms—with similar success rates. “Everybody's institution has different local climates and needs to adjust these things … It's nice to see that people are taking our drip, adapting it to their local environment, and having some success with it,” Dr. Goldberg noted.
And in 2004, the Yale group again updated its protocol following the publication of the first American Association of Clinical Endocrinologists' national guideline on inpatient diabetes and metabolic control (Endocr. Pract. 2004;10:77–82) and the American Diabetes Association's technical review (Diabetes Care 2004;27:553–91). The blood glucose targets were lowered to 90–119 mg/dL and the IV bolus was increased by about 40% to gain more rapid control.
In 54 consecutive cardiothoracic ICU patients, mean blood glucose levels were another 12–13 mg/dL lower on average with the new protocol and with no concomitant increases in hypoglycemia. Similarly, mean glucose level was 118 mg/dL among 47 consecutive medical ICU patients receiving 63 drips. With the old protocol, levels averaged 123 mg/dL. The new protocol halved to 4.5 hours the median time to reach a glucose level below 140 mg/dL (the old target).
These results would have been impossible without “buy in” from the nursing staff, Dr. Goldberg emphasized. “The ICU nurses are the ones who are doing this. You have to recognize that up front.”
A major barrier still to be overcome is the long-held fear of hypoglycemia. Many hospital personnel believe that levels of 150–200 mg/dL are “normal” and that anything below 100 mg/dL is cause for concern. “There is a 'culture of hyperglycemia,' with a fear of hypoglycemia, or even of low normal,” he said.
To address these concerns, inservice training at Yale consists of 35 minutes addressing the “why” of the protocol and just 10 minutes for the “how.” The trainers review the published data and reinforce the message that most hypoglycemic episodes are benign and treatable.
It's also important to acknowledge to the nursing staff that the infusions will cause them extra work, Dr. Goldberg said. Some of the impact can be minimized with efficient use of ancillary staff, additional glucose meters, and use of lines in place for other reasons to sample venous or arterial blood for glucose measures.
Continuous glucose monitoring systems—currently approved for use only in diabetic outpatients—might also prove useful in the ICU setting. In a preliminary study, the Yale group found good correlation between values obtained with Medtronic Minimed's CGMS system and capillary glucose levels in 22 medical ICU patients (Diabetes Technol. Ther. 2004;6:339–47).