Inpatient insulin: Secrets to smoother switches

Irl B. Hirsch, MD, MACP, offers tips for moving patients to formulary options.


Prescribing for a diabetic inpatient whose normal insulin regimen isn't available in the hospital is tricky even for the experts, Irl B. Hirsch, MD, MACP, told attendees at the American Diabetes Association (ADA) 2020 Scientific Sessions, held virtually in June.

Image by Getty Images
Image by Getty Images

As evidence, see survey results that Dr. Hirsch and Boris Drazin, MD, PhD, published in the June Journal of Clinical Endocrinology and Metabolism. The survey asked clinicians how they'd handle hypothetical scenarios in which patients on insulin degludec underwent hospital admission or a procedure.

For a patient on degludec and semaglutide who was scheduled for a colonoscopy the next day, endocrinologists gave a wide range of responses. “Only 10 of the 30 would keep the degludec dose the same the night before the procedure. Some of them would stop it, some of them would give half the dose,” said Dr. Hirsch, a professor of medicine at the University of Washington in Seattle.

“But that wasn't the surprising part,” he added. Three of the 30 subspecialists also wanted to hold the patient's semaglutide. “How were you going to hold a weekly drug if it's not the day that the drug is given?” he asked. “We have a lot of education to do, I think it's clear.”

Degludec is a particularly challenging drug, but only one of many that may pose problems for inpatient clinicians, according to Dr. Hirsch. “We have all of these drugs, and my anecdotal observation is that the people who are caring for these folks in the hospital, whether they're residents with attendings watching them on the internal medicine service or whether they are hospitalists, there are all these subtleties from new drugs that I don't think most people appreciate.”

Available varieties of insulin in the outpatient setting include human insulin in various concentrations, rapid-acting analogs, premixed analogs, pulmonary inhaled insulin, and an increasing number of basal insulins, including some combined with glucagon-like peptide-1 (GLP-1) receptor agonists.

However, when patients are admitted to the University of Washington, and most hospitals, the insulin options are much more limited. Dr. Hirsch's formulary choices are human regular insulin, insulin lispro, U500 regular human insulin, neutral protamine Hagedorn (NPH), premixed NPH and regular at a 70/30 ratio, and insulin glargine. “What happens when these people with all of these different insulin and insulin combinations come into the hospital and it's not on our formulary?” he asked.

Dr. Hirsch's talk tackled some of the most complicated versions of this question, starting with insulin degludec, which he noted has been an increasingly popular insulin globally, although not yet domestically. “For a variety of reasons, including the insulin cap pricing that we are seeing and the recent change from CMS about lower costs of insulin starting next year, I actually think the U.S. use of degludec will go up,” he said.

Thus, inpatient clinicians may increasingly have to switch patients from degludec to an available insulin, despite a lack of evidence on how best to do so. “We have, to date, no data to assist us in transitioning patients from newer insulins to formulary insulins, which could potentially result in major safety concerns,” said Dr. Hirsch.

There are at least some data on how the pharmacokinetics of degludec compare to those of glargine. According to research published by Clinical Pharmacokinetics in 2014, “The half-life of degludec is about twice that of insulin glargine. On average, it's about 25 hours for degludec and 12 hours or so for glargine,” he said.

The extended half-life has to be considered when switching a patient off degludec. “One can't just replace glargine or for that matter, NPH, for degludec when patients are admitted to the hospital,” said Dr. Hirsch.

His ideal solution to this dilemma would be for patients to bring their own degludec. “That doesn't happen much, but for patients who come into the hospital a lot, that's actually what they do, because this transition is usually not pretty,” he said.

The ugliness results from the risk that patients switched immediately to a normal dose of glargine will have hypoglycemia from insulin stacking, or the opposite problem, that holding the glargine and waiting for the degludec to leave their systems will lead to hyperglycemia, he said.

“To me the best solution, although not perfect, appears to be to give half that dose of glargine at the usual time on hospital day 1 and then the full dose on hospital day 2,” advised Dr. Hirsch.

The long half-life of degludec is also why the many survey respondents who said they'd change or skip that drug the night before a colonoscopy were wrong. “By the next morning, the insulin will be about the same if you withhold the dose, so making any adjustment in the degludec makes no sense at all,” he said.

A recent study by Dr. Hirsch and his colleague Jing Chao, MD, offers additional evidence of how few inpatient clinicians are familiar with this issue. Their retrospective analysis of 38 charts of patients from their hospital who were switched from degludec to glargine was presented as a poster at the ADA meeting.

“It was so interesting what we found—absolutely no consistency on how the transition was made. We saw everything. The glargine was withheld, it was given as a full dose, and everything in between, on hospital day 1,” said Dr. Hirsch.

The good news is that these inconsistencies rarely caused significant problems for patients with type 2 diabetes, who made up the majority of the studied population. Two of five patients with type 1 diabetes developed hypoglycemia, however. “Perhaps insulin stacking is only a safety issue in the type 1s. The number is small, but that's what we found,” he said.

The consequences could be more dire in next tricky situation Dr. Hirsch tackled—a patient admitted on pulmonary inhaled insulin, which is increasingly popular in the U.S. “In my mind, and I don't know where I heard this, the conversion is two-to-one, that is, for every two units of the pulmonary inhaled insulin, you need one unit of a rapid-acting analog,” he said.

This estimate turns out to be incorrect, however, based on a study that compared inhaled and injected prandial insulin. Patients had similar HbA1c levels on 110 units of inhaled insulin per day and on 26 units of insulin aspart per day, according to results published by Diabetes Care in 2015.

“That's a four-to-one ratio. My conclusion from this is that if one of these patients comes into the hospital, you only want to give about a quarter of the dose of the rapid-acting insulin to start,” said Dr. Hirsch. “And gosh, don't give that full dose of 110 units a day, because I think if you do, there's going to be a problem.”

Problems can also arise in the switch from U300 insulin to U100. The former lasts longer but is not as strong. “A very important point from the package insert: Patients treated with Toujeo, the U300 glargine, used 17.5% more basal insulin than patients treated with the U100 insulin, which is Lantus,” said Dr. Hirsch. This was true in both type 1 and type 2 diabetes and should guide inpatient practice, he advised.

“My interpretation is for an acutely sick patient taking U300 at home, transitioning to U100 would usually mean reducing the dose by 15%. We increase it by 15% when we go from U100 to U300,” Dr. Hirsch said.

However, it's also important to consider the particulars of a patient's acute illness, including the possibility of increased insulin resistance. “When somebody's admitted in the hospital, I want to have as much flexibility for treating their hyperglycemia as possible, and the reason for that is that their insulin requirements are going to be different, whether they are well controlled or they are poorly controlled. Whatever was happening on the outside is not going to be happening on the inside,” he said.

To enable such flexibility, he usually switches patients who are on outpatient regimens of premixed insulin to glargine and lispro. Dr. Hirsch also noted that lispro and aspart seem fairly equivalent, based on a 2019 study in Diabetes, Obesity and Metabolism. “In the hospital, one can substitute any of the rapid-acting analogs and the fast-acting aspart with each other,” he said.

As for the insulins that are combined with GLP-1 receptor agonists, Dr. Hirsch favors replacing them during hospitalization. “The reason why I don't like the GLP-1s in the hospital is, depending on why [patients] are in the hospital, they may be receiving narcotics. Those narcotics are going to slow gastric emptying on top of the GLP-1s, which are already slowing gastric emptying. I don't like to do that.”

On this and all of the other dilemmas he addressed, there is not enough evidence to definitively support any choice. “All of these insulins were initially tested and FDA-approved for outpatient therapy. To my knowledge, none of these were actually pretested with their registration trials for use in the hospital,” said Dr. Hirsch. “Please understand that I'm giving you these recommendations with no data.”

There is still one way for inpatient clinicians to get good data to guide insulin management, albeit with an n of 1. “All of the safety concerns that there may be with any of these issues, you can always get around with frequent testing, whether it is frequent fingerstick glucose testing or [continuous glucose monitoring],” he said. “With the plethora of new insulins and more coming, transitioning in the hospital to formulary insulins is a challenge which will only grow.”