Published in the January 2006 issue of Today’s Hospitalist
While the patient was admitted to the hospital with chest pain and ended up needing bypass surgery, the real trouble came in managing her diabetes.
The patient had a long history of type 1 diabetes, but the condition had always been well-controlled, in part because she was very diligent about her diabetes care. The surgery went smoothly, but when she was discharged from the ICU to the wards on sliding scale insulin only, she experienced ketoacidosis, lengthening her hospital stay by several days.
According to Thomas Donner, MD, director of the Joslin Diabetes Center at the University of Maryland School of Medicine, the above case, which involved one of his patients, illustrates a simple but critical point: For many diabetics admitted to the hospital, sliding scale insulin doesn’t do an adequate job of controlling glucose.
In a presentation on inpatient diabetes management at the Fall 2005 Hospitalist CME Series, Dr. Donner said that countless studies have shown that poorly controlled glucose is associated with many problems, from infections to mortality. He also pointed to research that has found that sliding scale insulin, an approach that is still common at many hospitals, leads to more hyperglycemia and hypoglycemia.
During the meeting, which was held in cooperation with Today’s Hospitalist magazine, Dr. Donner outlined some easy-to-use strategies to implement a standardized glucose regimen that will better control your inpatients’ glucose levels.
The trouble with sliding scale insulin
Dr. Donner acknowledged that tightly controlling the glucose levels of hospitalized patients is not always easy. He noted that some of the more obvious challenges include high levels of cortisol in inpatients, unpredictable eating times and little to no physical activity.
“When you think of managing diabetes, you think it’s really hard to do, and it can be,” he told the hospitalists at the meeting. “There are many obstacles to achieving good blood sugar control in these patients.”
Nevertheless, he said that the approach used at many hospitals is woefully inadequate. “If a patient comes into the hospital,” Dr. Donner said, “often times the housestaff will write sliding scale insulin orders regardless of whether the patient is on insulin or not.”
Giving a type 1 diabetic in the hospital sliding scale insulin won’t work for one basic reason. “Sliding scale insulin is retrospective in nature,” Dr. Donner explained. “You are treating with insulin after the fact. You are treating hyperglycemia after it’s already occurred.”
He cited a study of sliding scale insulin published in a 1997 issue of the Archives of Internal Medicine that proves his point. Investigators compared two groups of patients admitted to the hospital with a secondary diagnosis of diabetes. One group received sliding scale insulin, while the other group received a set dose of long-acting insulin with no sliding scale insulin.
Dr. Donner said that the fixed amount of long-acting insulin strategy actually outperformed the use of sliding scale insulin alone. The study found that patients who received sliding scale insulin only had higher rates of hypoglycemia and a threefold increased risk of hyperglycemia.
The building blocks of a standardized approach
Because sliding scale insulin produces such mixed results in the inpatient setting, Dr. Donner urged hospitalists to take another approach: a standardized subcutaneous insulin regimen. Why choose insulin instead of agents patients take at home? Dr. Donner said that most of those agents aren’t appropriate in the inpatient setting.
A widely used diabetes drug like metformin, for example, can cause life-threatening problems like lactic acidosis in patients predisposed to the condition, so it is rarely used in the hospital. And better-controlled patients taking sulfonylureas can develop hypoglycemia in the hospital when meals are withheld.
Besides, he added, insulin gives you much more flexibility than other agents. “You can adjust the dosage and type of insulin depending on whether the patient is getting IV dextrose,” Dr. Donner said, “whether they are on TPN, whether they are getting enteral feeds or they are getting nutritional supplements.”
To help hospitalists create a standardized insulin regimen, Dr. Donner started by reviewing the three basic types of insulin, which he said are the building blocks of a successful glucose control strategy. By choosing a combination of the following three types of insulin, hospitalists can tightly manage their patients’ glucose levels:
“¢ Basal insulin. This is the patient’s baseline level of insulin that is available all day long. Basal insulin gives the patient enough insulin to suppress hepatic glucose output, and it keeps the body from becoming hyperglycemic and ketoacidotic when patients are not eating.
“¢ Prandial or nutritional insulin. This type of insulin is administered to cover any nutrition the patient is receiving, whether it’s TPN, enteral feeds, meals or snacks.
“¢ Supplemental insulin. This is rapid-acting insulin and should be given to correct hyperglycemia detected through finger sticks.
Mimicking a normal metabolism
To choose an insulin to manage your inpatients’ diabetes, Dr. Donner said, try to picture the normal metabolism of nondiabetic patients.
If you looked at their glucose levels on a chart, you would see that they remain relatively fl at throughout the day. At meal times, however, those insulin levels spike up. It’s at this point that most type 1 diabetics need prandial or nutritional insulin to keep from becoming hyperglycemic.
While this may seem like basic physiology you learned in medical school, Dr. Donner said it’s what you want to mimic when writing orders for insulin.
You want to start by giving patients a steady level of basal insulin. You should typically provide half of the patient’s total daily insulin as basal insulin to control glucose levels when the patient is not eating.
“That’s a good take home point,” Dr. Donner said, “because when we talk about using basal and prandial insulin, we tend to use about 50 percent of the total dose as the basal insulin.” The other 50 percent of daily insulin should be divided up and given in prandial doses around meals and snack times for all type 1 patients, and for type 2 patients who have post-prandial hyperglycemia on basal insulin alone.
Choosing the right insulin
To provide the right amounts of basal and prandial insulin, you need to choose from the available therapies by examining their characteristics.
“When we try to mimic normal insulin physiology,” Dr. Donner said, “we look to replace basal insulin with something that is relatively fl at. For mealtime insulin, we look for a relatively short-acting insulin.”
When it comes to selecting an agent for basal insulin, for example, you can use either a long-acting insulin such as insulin glargine, or you can use intermediate-acting NPH insulin, Dr. Donner said. Insulin glargine is the longest acting insulin available, he noted, and it gives patients a continuous level of insulin throughout the day, lasting about 24 hours. “Of all of these types of insulin,” Dr. Donner noted, “insulin glargine is the best pure basal insulin currently available.”
NPH insulin doesn’t take effect until one to two hours after it’s been given and it peaks about four to eight hours later. It lasts for up to 20 hours, making it both a basal and prandial insulin.
Rapid-acting insulins like lispro, aspart and glulisine peak after about one hour and last for about four hours. As a result, Dr. Donner said, these are ideal agents to use before mealtime to cover the patient’s insulin needs during the meal.
Long- and rapid-acting insulin
Regular insulin is also short acting, but it takes about 30 minutes to take effect, and it can last for 10 to 12 hours if you use fairly large doses. As a result, Dr. Donner said, it ‘s not an ideal agent to control post-prandial glucose.
If you’re giving insulin to patients before meals, Dr. Donner suggested that rapid-acting insulins are the way to go because they can be given immediately before meals. Regular insulin, on the other hand, must be administered about 30 minutes before the patient eats.
The best way to mimic normal physiology is to give patients a combination of several types of insulin. A common strategy, he said, is to give a single injection of insulin glargine to provide basal insulin and a rapid-acting insulin analog to cover meals.
While supplemental insulin is one way to account for short-term changes in patients’ glucose levels, Dr. Donner warned against using them as a long-term strategy. “After you have written insulin orders,” he said, “you need to recognize that a patient’s insulin requirements are going to change as the severity of the illness goes up or down, or if their food intake changes, so you should be looking at the finger stick glucose levels on a daily basis. If the patient is requiring supplemental insulin over and over again, increase the amount of basal insulin. If you start seeing a trend toward hypoglycemia, it’s time to back off on the basal or prandial insulin.”
Because anticipating and working around patients’ meal times is one of the big challenges in controlling inpatient glucose, Dr. Donner reviewed a number of scenarios that can challenge physicians’ efforts to control glucose levels.
“¢ Patients who are eating regularly. Dr. Donner said that when treating patients who are eating regularly, ask them to try to consume a consistent number of carbohydrates.
“If you have somebody who is fluctuating quite a bit in terms of their glucose excursions,” Dr. Donner said, “it’s probably because they are eating differently from meal to meal. If you have somebody who is eating a big plate of pasta for lunch one day and they have a salad the next day, their insulin needs are going to be very different.”
He suggested explaining to the patient how the consumption of carbohydrates can cause fluctuations in glucose levels. One option is to ask a nutritionist to counsel patients on how to keep their carbohydrates consistent from day to day.
Dr. Donner also noted that some hospitals now prompt patients to select to a carbohydrate number (1 through 5, for example) at each meal in advance. The goal is to make sure that they receive roughly the same amount of carbohydrates from day to day.
“¢ Patients who are not eating. Dr. Donner said that if you’ve got the right dosage of basal insulin, your patients’ glucose level should not change, as long as they’re not eating and their stress level hasn’t changed.
“If you have patients who are intermittently eating food,” he added, “either because their appetite is variable or they are going for tests frequently, keep the basal insulin going and give rapid acting insulin with meals. If they’re not eating a meal, they don’t get rapid-acting insulin.”
“¢ Patients receiving total parenteral nutrition. Dr. Donner said that TPN commonly leads to hyperglycemia in patients with type 1 diabetes because their insulin dosage requirements can rise as much as 20 units over 24 hours. He also noted that just over three-quarters of patients who have type 2 diabetes but are not taking insulin at admission will need to be put on insulin when they receive TPN.
To deliver insulin to these patients, Dr. Donner suggested putting it directly into the bag, optimally after 24 hours of IV insulin to better assess the insulin requirement.
“¢ Patients taking glucocoritcoids. Glucocoritcoids increase the incidence of postprandial hyperglycemia more than fasting hyperglycemia. That’s why Dr. Donner said that IV insulin is the best way to control glucose in patients taking these drugs. “If patients are on steroids and are getting a transplant or have a serious infection,” Dr. Donner said, “IV insulin is the best way to get their glucose levels down quickly.”
He added that particularly in type 2 patients who didn’t need insulin before being hospitalized, you may need to give higher doses of prandial insulin because glucocorticoids boost postprandial hyperglycemia. “Frequent glucose monitoring in these patients is critical,” he added. “You have to change your dosage based on what the glucose measurements show.”
Switching from IV to subcutaneous insulin
Finally, when it comes to managing inpatients with diabetes, Dr. Donner said that problems can develop when switching from IV to subcutaneous forms of insulin.
Consider the patient treated for DKA whose glucose levels have returned to normal and looks well, but is still in the ICU on an IV insulin drip. That bed is suddenly needed by someone else, so the patient is moved to the wards.
All too often, Dr. Donner said, an order is written for sliding scale insulin until the long-acting insulin can be started the next morning. Not surprisingly, the patient wakes up the next day with DKA.
“You need to have subcutaneous insulin on board and especially a long acting insulin on board before the insulin drip is discontinued,” Dr. Donner explained. “IV insulin has a very short half life, so you should only discontinue the IV drop one hour after a combined rapid- and long-acting insulin, or two hours after an intermediate or long acting insulin has been injected.”
What if you typically give the patient insulin glargine at bedtime, but you have to stop the insulin drip at noon? Dr. Donner suggested giving a rapid-acting insulin to cover lunch, and then giving a small dose of NPH insulin as a basal insulin to last until the insulin glargine is given that evening.
Edward Doyle is Editor of Today’s Hospitalist.
Tips to calculate the insulin dosage of your inpatient
Your newly admitted patient has type 2 diabetes, and you’ve decided to switch from his oral medications to insulin. What’s the best way to determine the right dose?
If you try to figure out the patient’s dose by using trial and error, you’ll likely spend days adjusting the insulin level up and down. But there is another way.
Thomas Donner, MD, director of the Joslin Diabetes Center at the University of Maryland School of Medicine, said that whenever possible, he prefers to give these patients an IV insulin drip for 24 hours to determine the hourly drip rate. He then converts that figure into a 24-hour amount of insulin that can be added to the TPN bag on day two of the patient’s stay.
Dr. Donner acknowledged that it’s not always possible to take this approach, but he said research has shown that it works. He explained that one study showed that if the last few hours of a patient’s IV insulin infusion rate are steady, you can use those levels to calculate the patient’s 24-hour requirement.
Remember, he added, that if the patient is going to leave the ICU, you need to reduce that figure by 20 percent.
“When patients come out of the ICU, they are less stressed and they tend to need less basal insulin,” Dr. Donner explained. And if the patient is eating regularly, cut the figure in half. Give the patient the other half of the medication through prandial forms of insulin.
Once you’ve calculated the dosage, he said, watch the finger stick glucose levels and adjust the insulin doses the next day based on those results.