Intensity of insulin therapy may get all the press, but there are plenty of other dilemmas in inpatient glycemic control, as attendees at one Hospital Medicine 2011 session learned.
Endocrinologist John H. MacIndoe, MD, advised hospitalists on how to identify patients at risk of hyperglycemia, transition them between glucose control regimens during hospitalization, and discharge them with a plan that's likely to be followed.
The process begins at admission, said Dr. MacIndoe, who is the former director of endocrinology at HealthPartners Medical Group in Minnesota. “In any patient showing hyperglycemia, an A1c needs to be drawn as soon as possible after admission unless there is one on the charts in the last 30 days. It's also important from the very get-go to identify those patients that have financial or social barriers that are going to complicate their subsequent outpatient diabetes management.”
Potential social issues include poor home support, no transportation to a pharmacy or doctor's office, and drug and/or alcohol problems. Financial barriers may be posed by no insurance or a plan with a high deductible, including the Medicare doughnut hole.
“If you can get questions [about these issues] as part of your routine social history, you're likely to pick up on them early,” said Dr. MacIndoe. “If you get these problems out on the table very early, I think you've got a much better chance of identifying solutions.”
Solutions could involve different medications for the patient at discharge (more on that later) and the assistance of other clinicians. “Once you've identified the people with those barriers, it's critical to get diabetes education and social services involved very early,” Dr. MacIndoe said.
Oral medications and insulin
Another early-onset dilemma in glucose management is the question of whether the patient should continue taking oral medications during hospitalization. While there are insufficient data on this issue, Dr. MacIndoe recommends that oral agents only be continued in patients who are expected to have a very short stay with no disruption in nutrition, and selectively in those who are not critically ill, are eating regular meals, and have no contraindications. When the medications have been discontinued for minimally stressful surgeries, they can often be restored as soon as the next morning, he said.
“In our total knee replacement patients, it's working beautifully,” he said.
But patients with diagnosed diabetes and related medications are only one of the hyperglycemic groups in the hospital. Hospitalists will also have to deal with patients who develop hyperglycemia temporarily as a result of illness or treatment, and those with undiagnosed diabetes.
“Regardless of the etiology of their hyperglycemia, they all need to be treated aggressively to maintain their blood sugar in the target range. That usually means the use of insulin, either intravenous switching to subcutaneous, or subcutaneous,” said Dr. MacIndoe.
As with the oral meds, evidence is also lacking on the best way to transition patients from intravenous to subcutaneous insulin regimens, he said. “There are several protocols that have been published with supportive data but at this point, no one's done a head-to-head trial comparing them with one another, so we really don't know which one is the best.”
However, there are a couple of commonly accepted principles for transitioning from IV to subcutaneous, including the combined use of bolus (usually a rapid-acting analogue) and basal dosing. Also, “You want to be sure that the patient's blood sugar is stable for at least six hours before you make the transition,” Dr. MacIndoe said.
If the patient has been NPO and will remain so after transition, a simple way to calculate the appropriate subcutaneous dose is to calculate the patient's average hourly intravenous insulin rate and multiply that by 20 (because the subcutaneous dose should be 80% of the daily intravenous dose). The resulting amount should be given as a single daily dose of subcutaneous basal insulin, which Dr. MacIndoe usually provides as glargine. That patient doesn't need any other scheduled insulin, but additional sliding scale short-acting insulin should be given as needed every four hours.
Factoring in nutrition
Patients who eat complicate the process, however. If they were already on enteral feeding, then half of their IV insulin was covering nutrition and half was covering their basal needs.
“Generally speaking, 50% of the insulin that a patient takes throughout the day covers their basal insulin requirement, that is their hepatic glucose output, and the other 50% is divided up throughout the day to cover their nutrition,” said Dr. MacIndoe. Thus the new total daily requirement should be divided equally between basal (glargine once daily) and nutritional coverage. If the nutrition plan is to continue continuous enteral feeding, Dr. MacIndoe prefers to cover that by dividing the nutritional insulin by three and giving one-third every eight hours. Additional short acting insulin is then given every four hours as needed for blood sugars over 150 mg/dL.
If the patient is moving from enteral feeding to meals, the nutrition should come in the form of a constant (as in consistent, not total) carbohydrate diet. “Most folks order an ‘ADA diabetes diet,’ but there has not been one of those for years. Most nutritionists know that and so when they see ‘diabetic diet,’ in almost every institution, they transfer the patient to the constant carbohydrate diet, which is essential so you can predict the amount of insulin they should get with every meal,” he said.
Once the patient is on a constant carbohydrate diet, the nutritional insulin component is again divided by three and one-third is given as short-acting insulin with each meal. Since the scheduled nutritional dose only covers the carbohydrates in the meal, an additional sliding scale is added to the nutritional dose if the pre-meal blood sugar indicates a blood sugar higher than 150 mg/dL.
But a patient's failure to eat the ordered meal will complicate those predictions. “One of the toughest things to gauge in a patient, particularly right after transition, is whether or not they're going to start to eat, and if so, how quickly,” said Dr. MacIndoe. One way to deal with that uncertainty is this algorithm:
- Patients get their blood sugar checked at mealtime, as ordered.
- If that blood sugar is above target, patients receive whatever sliding scale is called for, even though they may not be hungry. (“This is a difficult thing for the nurses to understand, so it's important to go over that,” he added.)
- Nutritional doses should be held until patients are finished eating. Those who have completed more than half their meal receive a full dose of insulin. Those who have eaten less than half the meal receive half a dose.
Another new trick to teach the nurses: At the start of the transition from IV to subcutaneous, give the patient the basal dose and a rapid-acting bolus equal to 10% of the basal dose. “Because it kicks in so quickly, there will not be a lag, and blood sugar won't rise and you can discontinue the insulin drip right away, which will make the ICU staff very happy,” said Dr. MacIndoe.
He also offered some advice on sliding scale dosing. “At our institution, we're introducing a choice of three sliding scales, since one size does really not fit all given the differences in insulin resistance from patient to patient.”
According to Dr. MacIndoe's protocol, patients whose basal insulin dose is under 40 units get one unit of rapid-acting analogue for every 50 mg/dL that their blood sugar is over 150 mg/dL, while patients getting 40 to 80 units get twice that, and those taking over 80 get three times it. Rapid-acting analogues can also be used to cover patients' snacks, at one unit per 10 grams of carbohydrate.
How the initial A1c can help
And remember that A1c taken at admission? One way it can be used is in the transition from IV insulin. In a small trial published in the October 2010 Journal of Hospital Medicine, if a patient's A1c at admission was less than 6%, the protocol called for no basal subcutaneous insulin after transition from IV insulin and simply followed the patient with sliding scale. Dr. MacIndoe noted that although the study had good results, he hasn't had a chance to try this protocol yet, and it might not work in all patient populations.
He does use the A1c to decide how to treat patients headed for discharge.
“After the acute phase, their need for insulin starts to drop,” he said. “You're supposed to come up with a crystal ball to figure out what all this means in terms of what they need to go home on. It turns out you have a crystal ball in your hands. Almost all of the decisions that you need to make in terms of glycemic control are based on that very important thing you did at admission, that hemoglobin A1c.”
Here's how he uses the admission A1c to determine discharge instructions:
- Less than 5.7%. These patients probably don't have diabetes, so take them off insulin and check their blood sugar for 24 hours, and then “don't worry about it.”
- 5.8% to 6.5%. These patients meet the criteria for pre-diabetes but they should be able to get off insulin. Send them home with some advice, suggested Dr. MacIndoe: “You can prevent or at least postpone getting real diabetes if you lose weight, get exercise.”
- 6.6% to 6.9%. These patients have well-controlled diabetes. “If they're newly diagnosed, these are patients who need to understand how to improve their lifestyle to maintain control,” Dr. MacIndoe said. Adding metformin is also an option. “Obviously, if there's a patient who has diabetes and comes in with this kind of A1c, you don't want to rock the boat. Send them home on what they came in on.”
- 7% to 10%. These patients are not well controlled. “If newly diagnosed, you really should send them home on something, usually an oral agent, and preferably an oral agent that is not going to cause hypoglycemia, such as metformin or a thiazolidinedione. If this is a known diabetic who came in on poor control, you now have the opportunity to get them back on the path of righteousness by increasing their regimen,” said Dr. MacIndoe.
- More than 10%. These are the only patients that Dr. MacIndoe discharges with a new prescription for insulin. “Most people don't have the mindset in this setting to absorb that information [about how to use insulin]. It's critical that patients with this kind of high A1c see the diabetes educator very quickly after discharge or as soon as they're able to comprehend,” he said. In general, diabetes educators can be a good follow-up resource for patients because they are often able to see a patient more quickly than a physician and will focus on diabetes rather than any comorbidities.
If a patient requires new medication and was identified at admission as having financial barriers, then cost may influence prescribing decisions. “They're likely not going to tell you they can't afford the medication. They simply won't get the script filled and in time find themselves back in the emergency room for readmission,” said Dr. MacIndoe. “Most of us haven't got a clue about the street costs of diabetic medications.”
To remedy that situation, he offered a quick summary of diabetic medication and supply costs, starting with test strips, which typically cost $1.25 apiece. “If you put that patient on q.i.d. blood sugars, that's $35 out of pocket each week, $1,700 each year.” An alternative is the True Result test strip from Home Diagnostics. “It's 50% of the cost of all the other ones that are out there and they are accurate,” Dr. MacIndoe said.
Approximate monthly out-of-pocket costs for some commonly used brand-name drugs are:
- meglitinide (Prandin), $187
- pioglitazone (Actos), $236
- rosiglitazone (Avandia), $130
- exenatide (Byetta), $280
- sitagliptin (Januvia), $200
“This is not cheap stuff,” said Dr. MacIndoe. “On the other hand, if you look at the generic metformin and sulfonylurea options that are out there, at discount pharmacies you'll find that they can get month-long prescriptions for about $4 or three months for $10 for all of these agents.” Such drugs include metformin, glipizide, glyburide and glimepiride.
“Insulin is even more surprising when I look at the difference in pricing,” said Dr. MacIndoe. For analogue insulin, cost is about $135 per 10-mL vial or $200 for a five-pack of pens. “In contrast, good old NPH [neutral protamine Hagedorn] and regular insulin are about $25 to $35 a vial, although they're not available in pens,” he said. “Given that cost differential, I would suggest you strongly consider sending patients who have financial barriers home on twice daily NPH and regular.”
Younger physicians might be less familiar with the management of older insulins, so Dr. MacIndoe offered some advice. “That combination [NPH and regular insulin] will work if you remember one thing: The patient needs to eat regularly at mealtime and at snack time in order to stay ahead of the insulin they've given themselves,” he said.
Staying ahead of the challenges of insulin is, of course, also the goal for hospitalists, to whom Dr. MacIndoe offered a summary of his instructions in parting. “Get that [initial] A1c as close to admission as you can and act on any barriers that have to be dealt with. Design an affordable treatment plan that is both safe and effective. It's critical that patients be seen 7 to 14 days after discharge, particularly if they're on a new regimen.”