Dominate inpatient diabetes! In this part 1 of 2, we discuss how to handle oral hypoglycemics, how to choose an initial insulin regimen, the use of correction scale insulin monotherapy, how to titrate insulin, how to use NPH for steroid-induced hyperglycemia (check out this Twitter thread on safe dosing), and how to reconcile diabetes meds at discharge!
More inpatient diabetes! In this part 2 of 2, we cover diabetic ketoacidosis (DKA), euglycemic DKA, how to handle insulin pumps in the hospital, continuous glucose monitors, and how to transition patients off an insulin drip! Our guest is Dr. Dave Lieb (@dclieb), an internist, endocrinologist and the endocrinology program director from Eastern Virginia Medical School.
What will your patient’s discharge plan look like? What can they afford? What kind of education will they need? Regarding education, Dr. Lieb suggests that any opportunity to educate a patient and/or their family should be seized!
Dr. Lieb reminds us that context is key! If a patient is undergoing cardiac surgery, tight control may be important. Conversely, if the patient is at high risk for hypoglycemia, tight control may be deleterious. Patient-centered care is a priority when it comes to glucose control.
Medications that a patient takes routinely at home may introduce side effects or confound the overall treatment plan for a patient during their admission. Admissions can often be unpredictable regarding procedures or the general ebbs and flows of a patient’s course. Therefore, stopping home diabetes medications (oral and injectables) and placing the patient on an insulin-only regimen can often be the safest and most flexible means to manage their blood sugar during a hospitalization.
In some cases – for example a prolonged wait for disposition – a patient may be out of the acute phase of their hospitalization and restarting home medications / stopping insulin may be appropriate, on a case-by-case basis.
Remember: It is critically important that the discharge plan is explicit and clear in regards to what to restart, what to hold, and why any changes were made.
There are a lot of types of insulin out there and the landscape is ever-changing! Insulin helps you in three basic ways: Low-level insulin (i.e. basal) that suppresses hepatic glucose production between meals and overnight, mealtime insulin that covers the glucose-spike from a meal (ideally given 20 min prior to mealtime), and correctional insulin that accounts for high blood sugars (i.e. correctional insulin). Dr. Lieb reminds us that, in general, your hospital will have an insulin formulary that limits your choices. However, prior to discharge, the myriad of options can be mind-boggling! Use your hospital pharmacist to help you find the best discharge regimen for your patient (often based upon cost / insurance).
Insulin glargine is a commonly used basal insulin. It has a duration of action of approximately 24 hours. Alternatively, insulin detemir can be used for basal coverage and is particularly helpful in renal dysfunction given its half life of 12-20 hours. Rapid options include insulin lispro or insulin aspart. [Donner & Sarkar, 2019]
A note on rapid-acting insulin: Given the risks of hypoglycemia and the unpredictabilities of inpatient medicine, it may be reasonable (although not optimal) to dose mealtime insulin immediately when food is served or afterwards.
Dr. Lieb recommends a multi-factorial approach to determining insulin dosing. Is the patient well-controlled or not? How sick are they? What do they weigh? A lot of hospitals will have weight-based order sets that can set you up for success – although they may need a bit of tailoring. For example: a total daily dose of 0.2-0.5 u/kg divided into 0.2-0.3 u/kg for basal insulin and 0.075 u/kg for mealtime insulin [Umpierrez et al, 2012]. Dr. Lieb reminds us that you can always give more insulin, but you cannot take it away – being conservative upfront is a very reasonable approach!
A few other thoughts from Dr. Lieb: If a patient is on insulin at home, you can start from there and adjust as needed. You may consider decreasing home insulin by 15-20% in a patient who is eating poorly, which is what Dr. Lieb does in his own practice. Also, a simple way to determine dosing may be to use the 0.2-0.5 u/kg daily dosing to determine a total, and simply divide that in two yielding basal dosing and mealtime dosing (which would be divided in three to account for breakfast, lunch and dinner). Another consideration, per Dr. Lieb, is that if you are using insulin in a more tenuous patient or someone who is insulin naive, providing less mealtime insulin than initially calculated may be a prudent approach. For example, if half of the total daily insulin is 30u, normally that would be split into 10u with each meal, however, opting for 4-6u may be safer (i.e. less likely to cause hypoglycemia) in certain circumstances.
What about bedtime insulin? While this is common practice in some institutions, Dr. Lieb recommends against this. Instead, it is safer to watch these patients initially and determine if they would benefit from a dose of rapid acting insulin before bedtime. Instituting this by default can precipitate nocturnal hypoglycemia!
Dr. Lieb recommends against using correctional insulin monotherapy – sometimes called “sliding scale,” in most cases [Ambrus & O’Connor 2018]. Correctional insulin should be used to correct hyperglycemia, and then a portion of the total daily correctional can be added to the next day’s basal-bolus regimen to achieve better glycemic control. Prior to making changes, Dr. Lieb recommends talking to the patient and nursing staff to determine if there is an explanation for any blood sugar spikes (for example, an unexpected late night snack). Overall, the titration of insulin in the inpatient setting is dynamic and requires the medical team to be deliberate and collect collateral information while making adjustments. However, hypoglycemia requires decisive down-titration of insulin given that the risks of hypoglycemia are significant. Dr. Lieb might consider a 20% dose reduction in the setting of hypoglycemia.
Who may be a candidate for correctional insulin monotherapy? A patient with mild or well-controlled diabetes who is not very ill, correctional insulin with vigilant monitoring and titration may be reasonable.
Dr. Lieb recommends a multi-factorial approach to determining insulin dosing. Is the patient well-controlled or not? Historically, a goal of 80-110 mg/dL had been used by many in the early 2000s based upon two studies out of belgium: one in a surgical ICU and one in a medical ICU [Van de Berghe 2001, 2006]. These studies were criticized as they recruited patients from one center in Belgium. In 2009 the famous, multicenter NICE-SUGAR trial compared this intensive regimen to a less-intensive regimen which showed improved mortality and far-fewer episodes of severe hypoglycemia in the more liberalized arm, which has since given rise to the 140-180 mg/dL goal [Finfer et al 2009]. Overall, the data shows uncontrolled hyperglycemia and severe hypoglycemia can both lead to morbidity and mortality, thus, a rough/ballpark goal of 100-200 mg/dL is reasonable for almost all-comers.
The “diabetic diet” in most hospitals is a carbohydrate controlled meal, which theoretically makes it easier to dose insulin. However, patients may often eat additional carbohydrates that can make dosing difficult. Dr. Watto reminds us that a patient-centered approach may be needed to maintain a productive therapeutic relationship with a patient, given that sometimes the diabetic diet may be frustrating for a patient. Remember, many patients will not maintain the same “diabetic diet” at home and thus, this may create an unrealistic eating environment.
Dr. Lieb notes that most of the hyperglycemia seen in association in steroids is post prandial, however, the effect of steroids on a given patient is unpredictable. Given that steroids present a moving target, Dr. Lieb will often utilize a continuous insulin infusion in patients getting high doses of steroids, to try to maintain better glycemic control (stress-dose steroids, transplant rejection doses, for example).
One pearl that Dr. Lieb shares is the use of NPH insulin in conjunction with steroid dosing. The peak-effect of NPH seems to correlate nicely with the peak-hyperglycemia seen with steroid use, based upon the pharmacokinetics of NPH. [Atkinson 2016] An approach published in Endocrine Practice in 2009 recommends an NPH dose equal to the product of the patient’s weight in Kg and 1/100th of their steroid dose up to a max of 0.4 units NPH per Kg body weight [Clore & Thurby-Hay 2009]. For example, a 90 kg patient on 60mg of prednisone would get [90 x (40/100)] = 36 units of NPH, just to cover the steroid effect. Another more recent article recommends a max initial dose of 0.3 units NPH per Kg body weight (Aberer 2021). Dr. Lieb says it may be prudent to start with a lower dose of NPH if you are concerned for hypoglycemia. He posted this Twitter thread to clarify the point.
Some patients may need to go home on insulin. If this is likely, Dr. Lieb recommends early education per the diabetes education team, nursing and the medical team. This includes didactic education, as well as practical teaching on insulin pen use, etc. For example, someone on meal time and basal insulin who has been difficult to control will likely need insulin at home. Similarly, a patient on triple-therapy who was poorly controlled and required high doses of insulin as an inpatient will likely need to go home with insulin, and thus appropriate discharge planning is of paramount importance. In the appropriate patient, a combination of basal insulin and other diabetic therapies may both achieve the goal and permit fewer daily injections. Dr. Watto underlines the importance of tailoring the discharge plan to the patient, as the plan is only as good as the follow through.
DKA is defined as being the presence of acidosis (arterial pH < 7.30), in association with hyperglycemia (serum glucose > 250 mg/dL) along with urine and serum ketones [Kitabchi et al 2009]. It is further classified as mild, moderate or severe based on the degree of laboratory derangement and the patient’s mental status. Additionally, there exists a related condition – HHS which typically presents in patients with very high blood sugar (>600 mg/dL) in the absence of acidosis. Dr. Lieb states that a major discriminating factor is how much insulin is “around.” Specifically, he reminds us that many patients with HHS suffer primarily from dehydration and that a significant degree of the patient’s hyperglycemia can be managed by fluid resuscitation.
IV insulin therapy is a mainstay of DKA care. Therefore these patients often benefit from admission to an intensive care unit. These patients are generally very dehydrated and thus, need significant fluid resuscitation. Often, the dehydration (glucosuria causing an osmotic diuresis) is associated with hypokalemia which requires aggressive management. Dr. Lieb recommends avoiding normal saline to avoid the hyperchloremic acidosis that it can cause. From a pathophysiology standpoint, patients either have an absolute or relative insulin deficiency (i.e. ketosis prone type 2 diabetes). Due to the insulin deficiency, these patients have a rise in serum glucagon which causes free fatty acid conversion into ketone bodies.
Sometimes DKA occurs due to a lack of insulin (i.e. a patient fails to take sufficient insulin for whatever reason). However, often DKA is precipitated by infection or some other stressor such as an acute MI, a progressive wound or trauma. Determining the precipitating factor for DKA is just as important as treating it! Dr. Lieb cautions us against the major pitfall of assuming a patient’s presentation is due to “non-compliance.” In fact he advocates against the use of that term – as do we at The Curbsiders – since there is always a reason for treatment failure that can and should be identified. This is especially important in patients who are frequently admitted to the hospital with DKA. From a pathophysiology, it seems that many of these stressors share a common pathway – i.e. the production of pro-inflammatory cytokines – which leads to derangements of glycemic control [Palmero et al 2020].
Treatment is generally protocolized, and these protocols are determined at the hospital level. It is imperative that, despite such protocols, clinicians still need to be vigilant in case the protocol is not adequately treating the patient. Alternatively, some patients may have mild DKA (pH>7.25, serum bicarbonate of 15-18 mEq/L, glucose > 250 mg/dL) and do not need insulin drip management but instead could be managed via subcutaneous insulin. Dr. Lieb references this recent paper in the Journal of Clinical Endocrinology and Metabolism that can help clinicians triage DKA severity [Palmero et al 2020]. As a bonus, this article discusses DKA in the era of COVID-19! In systems where ICU resources may be limited, this approach provides a means to manage mild DKA in the general medicine wards.
In regards to treatment goals, Dr. Lieb discusses a few measurables. Some providers will measure serum ketones and look for a downtrend. Monitoring electrolytes (especially potassium and phosphate) is very important. Perhaps more important than labs, is the patient. Are they in pain or bedbound? Or are they feeling better and developing an appetite?
The transition from insulin drip to subcutaneous insulin may be the most critical part of a patient’s hospitalization. Dr. Lieb implores us to make sure the patient feels better and has an appetite prior to stopping the drip. Additionally, he recommends electrolyte derangements should be treated, the patient should not be acidotic, have a normal anion gap and have a serum bicarbonate > 18 mEq/L. When transitioning, Dr. Lieb explains that the patient should receive basal insulin while on the drip, with a goal of a one to two hour overlap, prior to cessation of the drip. Once the patient’s food has arrived, the drip can be turned off (assuming that overlap occurred!) and then the patient should eat and receive their first dose of meal time insulin with any correctional insulin if indicated.
How much insulin should the patient be on? Dr. Lieb recommends looking at the last 4-6 hours while the patient was “NPO” and on the drip. This number, extrapolated over 24 hours (i.e. x4 or x6) gives a rough estimate of the patient’s basal needs. Doubling that amount is a rough estimate of their daily needs, allowing the clinician to devise a basal bolus strategy tailored to the patient. Some clinicians will be more cautious and multiply the calculated total daily insulin dose by 0.8, but this is clinical-dependent.
Most patients with DKA will have a blood glucose > 250 mg/dL, however, 5% of patients with DKA will have a blood glucose that is either euglycemic or less than 200 mg/dL [Long et al 2021]. In euglycemic DKA, the pathophysiology differs from hyperglycemic DKA in that there is a combination of a relative insulin deficiency with a concurrent glucose deficiency. This could be seen in pregnant patients, malnourished patients, patients with liver disease or who suffer from alcohol use disorder, for example. A common scenario that clinicians need to be aware of relates to the use of SGLT-2 inhibitors. These agents work by increasing glucosuria to reduce serum blood sugar. A patient who, for example, has developed sepsis from an infection but has continued their SGLT-2 inhibitor, is set up for developing euglycemic DKA, thereby complicating their clinical course. Should a patient who develops euglycemic DKA stop their SGLT-2 inhibitor? In Dr. Lieb’s expert opinion, these drugs are effective and it may be reasonable to continue them, especially if the patient has a strong indication such as concomitant heart disease or renal failure. Any patient on an SGLT-2 inhibitor should be cautioned to hold the drug in certain circumstances, such as during illness, when fasting, or during the perioperative period (Goldenberg 2016).
From a treatment standpoint, while these patients may not need aggressive insulin therapy for hyperglycemia management, they do require aggressive hydration and electrolyte management. They still require insulin in order to close their anion gap and treat ketosis. Due to their euglycemia, they may require more dextrose to prevent hypoglycemia than those with hyperglycemic DKA. Overall, Dr. Lieb cautions providers against missing these patients who have a reassuring blood sugar but otherwise need the same level of vigilance as classic DKA patients.
Insulin pumps are great… except when they fail! Dr. Lieb suggests that we counsel any of our patients with insulin pumps to be on the lookout for symptoms of hyperglycemia. If a patient feels malaise, fatigue or other symptoms without an obvious cause, he recommends a finger stick to check blood sugar, followed by insulin injection if indicated, rather than trying to troubleshoot the pump. This is a good way to prevent symptomatic hyperglycemia from becoming DKA in patients with pumps. Pumps can experience technical failures, they may “disconnect” from a patient, or the tubing can fail – all of which can lead to hyperglycemia. Once any acute issues are dealt with, the pump itself can be evaluated/repaired/replaced. To avoid any confusion, Dr. Lieb recommends selecting a new infusion site as well, if possible.
Most patients do fine in the hospital with their own insulin pumps. It is important to make sure a patient who plans to use their own pump during a hospitalization, is awake enough and not too sick to manage their pump. Dr. Lieb mentions that for many patients, the insulin pump provides patients with some degree of control and autonomy that can be very reassuring and help fortify the therapeutic relationship. Thus, if possible, it may be beneficial to empower a patient to manage their blood sugar with their pump if they are able.
Like insulin pumps, these are becoming more and more common. They are (supposed to be) calibrated regularly with fingerstick checks and allow for far fewer “pokes.” The use of CGMs in the hospital is a new frontier as there are still questions about accuracy and safety. Dr. Lieb suggests that the potential for these monitors is very exciting, especially in their ability to alarm patients regarding significant changes in blood sugar / serve as an early warning indicator of blood glucose derangements. However, right now, a CGM should not be used entirely in lieu of hospital-grade fingerstick glucometers.
Listeners will develop a patient-centered approach to managing inpatient hyperglycemia using various tools in a variety of scenarios.
After listening to this episode, listeners will…
Dr. Lieb reports no relevant financial disclosures. The Curbsiders report no relevant financial disclosures.
Barelski A, Askin C, Lieb D, Williams PN, Watto MF. “#330 & #331 Inpatient Diabetes with Dr. David Lieb”. The Curbsiders Internal Medicine Podcast. http://thecurbsiders.com/episode-list 25 April, 2022.
Got feedback? Suggest a Curbsiders topic. Recommend a guest. Tell us what you think.
We love hearing from you.
Yes, you can now join our exclusive community of core faculty at Kashlak Memorial Hospital along with all the perks:
Close this notice to consent.