Published in the May 2009 issue of Today’s Hospitalist
The patient was receiving 12 to 13 units of IV insulin per hour post-surgery in the ICU before being sent to a regular bed with well-controlled blood sugars. A few days later, however, he was back in the ICU with an infection “and blood sugar readings spiking above 300.
The transition from IV insulin in the ICU to subcutaneous insulin on the wards is notoriously challenging for hospitalists and intensivists alike, but the above case got the attention of physicians at Methodist Hospital in Memphis, where it occurred.
“We have to develop something to bridge this transition,” says Christopher K. Finch, PharmD, Methodist’s manager of clinical pharmacy services and a critical care specialist. “Maybe we could have prevented the infection if we had kept the blood sugars under control.” Spurred by the case, Dr. Finch and his colleagues are now pilot-testing an IV sub-Q transition protocol.
Across the country, it seems, clinicians are searching for the holy grail of inpatient insulin therapy: formulas, protocols and other approaches to head off destructive swings of hypo- or hyperglycemia when patients leave the ICU. They are convinced that fine-tuning glycemic control will provide a big payoff in terms of reducing readmissions to the ICU and overall morbidity and mortality.
Not everyone, however, is completely on board with those efforts. Some hospitalists, for example, view the need for better glucose transitions as controversial, in part because of what they say is a lack of strong evidence favoring tight control in large groups of patients like nondiabetics.
Others see less of a problem continuing to use sliding scale insulin, while some physicians debate how tight control should be for patients on regular floors. Strict blood sugar control may be a laudible goal, they add, but the chaotic realities of modern hospitals make well-crafted insulin transitions less of a priority. The mixed reaction has led to a lukewarm reception for transitional glycemic control at some hospitals “and a variety of approaches.
The scope of the problem
Anecdotes abound about the challenge of transitioning patients from IV insulin to sub-Q injections. But until recently, there were few data on just how bad the problem was.
In the January 2009 issue of the Journal of Hospital Medicine, Dr. Finch and his colleagues at Methodist reviewed what happened to 65 patients who had been on intensive insulin protocols in the ICU and then transitioned to sub-Q insulin on the floor.
They found that only 38% had scheduled insulin ordered for them on the wards “and that blood sugars for all patients rose significantly. During their first 12 hours outside the ICU, mean blood glucose levels for the patients studied were 168Â±50 mg/dL. That was much higher than the 123Â±26 mg/dL mean blood glucose measured during the preceding 12 hours in the ICU.
That jump in readings comes as no surprise to Cheryl O’Malley, MD, a hospitalist and glycemic control expert at Banner Good Samaritan Medical Center in Phoenix. “If you aren’t doing something to standardize insulin use in your hospital, you’re likely causing excessive hypoglycemia and hyperglycemia,” Dr. O’Malley says. “If you ignore it in the hospital, your patients are going to have extremes, which have metabolic consequences.”
An internal review at her hospital a few years ago found that only 55% of blood glucose draws were in the normal range during the first 24 hours after patients were transitioned from IV insulin in the ICU to sub-Q injections on the floors. The hospital defines normal range as 70-150 mg/dL.
After instituting a pilot transition protocol in one unit recently, however, the rate of normal readings increased dramatically to between 65% and 70%.
Is it still OK to have more than a quarter of readings indicating hyperglycemia? “I don’t know,” Dr. O’Malley admits. “There aren’t a lot of data about how much this will benefit patients, but our basic goal now is to avoid extreme hypo- or hyperglycemia. And we are doing that with the protocol.”
Choose your formula
At Methodist, Dr. Finch says clinicians tally how much regular insulin patients required during the last 12 hours of insulin infusion.
“We then take 60% of that amount and use it as our twice daily NPH dose,” Dr. Finch says. “In addition to the NPH, a short-acting insulin is utilized around meals, if patients are eating.” While Dr. Finch has no data yet on outcomes, “we haven’t had any major adverse outcomes.”
The formula being worked up at Dr. O’Malley’s hospital is a little different: When a patient is ready to transition off IV insulin, the nurses collect the last seven drip rates, throw out the two highest, add the remaining five together and multiply by four to determine the total daily dose.
“We want to underestimate what their total daily dose is because you would rather err on the lower side,” Dr. O’Malley says.
That amount is likewise cut in half, with half then given as basal “usually once-a-day, long-acting glargine (Lantus) “and the remainder as fast-acting nutritional insulin with meals. According to Dr. O’Malley, the first injection of glargine is supposed to be given two hours before the drip is turned off. She points out that this protocol is used for both type 1 and type 2 diabetics without diabetic ketoacidosis (DKA).
The protocol also covers non-diabetics experiencing stress-induced hyperglycemia whose GTT rate over the previous seven hours was more than 1.5 units per hour or those whose hemoglobin A1c is more than 6.5%. “All others,” says Dr. O’Malley, “just go on sliding scale.”
At the University of California, San Diego (UCSD) Medical Center, hospitalists take an even narrower approach. “Hospitalists apply the transition protocol only to patients with a history of diabetes,” says UCSD hospitalist Diana Childers, MD, “or to those whose hemoglobin A1c levels are greater than 6.0%.”
The center’s physicians decided to not use the protocol with other patients with stress hyperglycemia after seeing the results of a study conducted by Dr. Childers and another UCSD hospitalist, Pedro Ramos, MD. The study, which was presented as an abstract at last year’s Society of Hospital Medicine meeting, concluded that these patients don’t benefit from a transitional protocol. Using correction doses on the floor, researchers found, seems to work just as well.
Dr. Ramos explains that those findings are important because physicians have serious questions about which patients need to be transitioned to scheduled insulin and which will do just fine without all the extra work.
“Our data show that patients who come in with stress hyperglycemia whose A1cs are less than 6 do very well on just a correction scale,” Dr. Ramos says. “They may need a couple of units of insulin here or there as their illness gets better, but they don’t need to be on scheduled insulin.”
The UCSD transition protocol, which is now being pilot-tested, calls for taking the average insulin infusion over the last six hours and multiplying that number by 20. Half of that total daily dose is then given as subcutaneous basal insulin. (UCSD uses Lantus.) The other half is nutritional insulin, given as Lispro with meals, or every six hours if the patient is on continuous tube feeds.
The need for daily adjustments
Since they started using the protocol, Drs. Ramos and Childers have found that blood sugar levels of intervention patients are significantly better in the first few days after a transition than that of control patients (a mean of 167 mg/dL on the first day vs. 216 mg/dL; 175 mg/dL on the second day vs. 215 mg/dL).
Just as importantly, the protocol hasn’t produced any increase in hypoglycemic events, and it has significantly decreased the percentage of patients with values greater than 300 mg/dL.
Dr. Ramos points out, however, that a major obstacle remains: nursing concerns. “Nurses were afraid of giving the long-acting insulin on top of the drip and causing hypoglycemia,” he explains, “so they weren’t giving it before they stopped the drip.” By contrast, he says, giving a subcutaneous injection two hours before insulin infusion is turned off makes a big difference in patients’ first six hours of blood sugar control.
Another stumbling block in all these protocols: Physicians must remember to revisit sub-Q insulin doses daily and adjust as needed. Many patients, particularly those who don’t have underlying diabetes, will see their supplemental insulin needs decrease as their health improves.
“Be prepared after they make that transition to look at them every day and make adjustments down if their sugars are under good control,” says Dr. O’Malley. “The transition is just one step of proactive glucose management that should continue throughout the entire hospitalization.”
Resistance to protocols
In the minds of many hospitalists, however, the problems related to a smooth insulin transition go deeper than just coming up with a straightforward formula. Kendall Rogers, MD, chief of hospital medicine at the University of New Mexico Health Sciences Center in Albuquerque, for instance, says that the effort to craft effective transitions reveals deeper fault lines between the ICU and the floor.
“We have our critical care units getting infusion protocols running well for their patient population, and then we have a separate process trying to improve basal-bolus usage on the floor,” Dr. Rogers says. “Really, the marriage between the two has been on our backburner.”
That lag, he says, comes in part from the incredible fear of hypoglycemia “particularly as patients coming out of an ICU start eating more “that prompts both physicians and nurses to want to err on the side of allowing patients to be hyperglycemic.
“We have had resistance in our ICU to having a formalized transition to a hospital-wide, sub-Q-insulin protocol,” Dr. Rogers explains. “Our critical care doctors want a very conservative switch “without the scheduled use of nutritional insulin, due to inconsistent nutritional intake “to ensure that patients aren’t having hypoglycemic episodes.”
Then there is the fact that transitional care is marked by so much change that some physicians worry that standardization will end up causing more problems than it will fix. Will every patient, they ask, be an exception?
“In this day and age of protocols and order sets and pathways, I don’t think there is a way to pigeonhole these kinds of patients,” says Michael Mena, MD, hospital medicine chief at Valley Medical Center in Seattle, a Sound Inpatient Physician practice. Currently, Valley Medical has no protocol to help with transitions. “Good glucose control is important,” says Dr. Mena, “but I think that the use of an algorithm could be complicated and difficult to implement.”
Other major barriers, he adds, are clinical priorities, patient throughput and resource utilization. And poor control is usually a side effect of the acute event that landed patients in the ICU, he points out, not of hyperglycemia. While he supports efforts to create an adaptable transfer protocol for glycemic control to improve patient outcomes, “unfortunately, glucose control becomes No. 2, 3 or 4 on our problem list for these complex, critically ill patients,” Dr. Mena says. “It all has to do with the global picture of your specific patients.”
Another challenge for many doctors when it comes to transitioning insulin is the fact that the evidence is a little thin; most of the benefits demonstrated so far relate to the critically ill. Even those data have come under increased scrutiny, with recent studies showing increased rates of hypoglycemia with tight glucose control in the ICU.
“The benefit of tight glucose control on the wards is still pretty theoretical at this point,” notes Jeffrey L. Schnipper, MD, MPH, director of clinical research for Brigham and Women’s Hospital’s academic hospitalist service.
Dr. Schnipper has been working on a number of projects to improve the quality of inpatient management of hyperglycemia. But complicating those efforts, he says, is the reality that patients in most hospitals are shipped out of the ICU before their glucose control is perfect. Why? Because those beds are needed for someone sicker.
“You move patients because they are not intubated or on pressors anymore, not because of anything to do with glucose control,” he points out. On the other side of that transition, he adds, patients are going to a unit with logistical staffing issues that prohibit the kind of frequent blood checks possible in the ICU.
“That’s what makes it complicated,” Dr. Schnipper says. He agrees that if you could keep patients on an insulin drip through their transfer to a noncritical care unit and then transition them to sub- Q insulin, you would probably get better control. “But we all know that the transition is usually determined by very practical issues” “like bed and staffing issues “”and not purely based on clinical issues.”
In his hospital, he says, an ideal proper transition rarely occurs. Instead, when a patient leaves the closed ICU, “chances are the intensivists have written orders for sliding scale insulin on the floor and I basically have to start over,” Dr. Schnipper says. Now that physiologic insulin advocates in his hospital are making headway both in the ICU with insulin drips and on the floor with sub-Q protocols, he says he is just starting to feel he can work toward a much-needed transition protocol.
Frankly, “you do the transition protocol after you fix the other aspects of glucose management,” says Dr. Schnipper. “If glucose control in the ICU is lousy and glucose control on the floor is lousy, who cares if your transition is great?”
Deborah Gesensway is a freelance health care writer based in Sierra Madre, Calif.