Principles of Glycemic Management in the Hospital

The following principles are the foundation of treating patients with glycemic management issues:

BG targets in non-critically ill patients. [I-C]

Pre-meal

100–140 mg/dL for typical situations

120–180 mg/dL for significant comorbid conditions, hypoglycemia unawareness

Random

≤180 mg/dL for typical situations

<200 mg/dL for limited life expectancy and terminal illness

Physiologic insulin treatment in the hospital. Insulin is the preferred drug for glycemic control in the hospital. Table 1 lists the currently available formulations of insulin.

Basal-bolus insulin regimens most closely approximate physiologic insulin release and are the preferred inpatient insulin regimens. A step-wise approach to prescribing a basal-bolus insulin regimen is shown in Figure 1.

Monitoring glucose in the hospital. All patients with T1DM, T2DM, cystic fibrosis-related diabetes, patients on high-dose steroids, or those with stress hyperglycemia must have their BG monitored with point-of-care BG checks before meals (QAC) and at bedtime (QHS), or every 6 hours if NPO. [I-D]

Hemoglobin A1c. A1c should be checked in all hospitalized patients with known diabetes mellitus who have not had a measurement in the last 3 months, [I-B] and when the diagnosis of diabetes is being considered.

Management of Specific Clinical Scenarios

Outlined below are operational steps to achieve glycemic control in specific clinical situations:

Stress hyperglycemia in patients without diabetes (Figure 2)

T2DM, taking only non-insulin agents prior to admission (Figure 3)

T2DM, taking basal-only insulin, or basal insulin plus non-insulin agents, prior to admission (Figure 5)

Taking basal-bolus insulin prior to admission (Figure 6)

Titrate insulin based on Table 2.

New diagnosis of diabetes made in the hospital (Figure 7)

Before discharge, provide diabetes management education focused on survival skills (Table 3). Provide follow-up in an outpatient comprehensive diabetes program.

For all patients with a new diagnosis of suspected T1DM, obtain Endocrine consultation. [I-D]

T1DM

Never hold basal insulin in a patient with T1DM.

For T1DM patients who count carbohydrates at home, continue meal insulin dosing based on carbohydrate ratios in the hospital (with modification of the ratio if needed). [I-D]

For T1DM patients on a sensitivity factor or custom correction scale at home, use this as the basis for a custom correction scale in the hospital. [II-D]

Patients using an insulin pump (continuous subcutaneous insulin infusion) at home.

Contact the Endocrinology Service for assistance with managing all insulin pump patients in the hospital. A formal consult is not required if patient is educated, insulin pump orders are in place, and BG is well managed.

Patients on insulin-pump therapy at home can continue to use their pumps in the hospital if they meet existing safety criteria. (At University of Michigan Health System (UMHS), see the Insulin Subcutaneous Infusion Pump Therapy Policy.)

Steroid-induced hyperglycemia see Table 5

Pregnancy and Peripartum. For antepartum glycemic management, see Table 6. For postpartum glycemic management, see Table 7.

Perioperative Diabetes Management

Preoperative assessment of glycemic control.

Check a preoperative A1c in all diabetes patients to assess control (if not done within the prior 3 months).

Check a preoperative A1c on patients without known diabetes if they meet the criteria in Figure 8.

Glycemic management in the preoperative period.

Table 8 shows recommendations for preoperative medication management.

Figure 10 shows recommendations for BG monitoring and control in the preoperative/holding area.

Figure 8 discusses when to consider delaying surgery due to hyperglycemia during the preoperative evaluation, and Figure 9 discusses the same for the day of surgery.

Patients who use an insulin pump for diabetes management should be instructed on preoperative management. (At UMHS see the Perioperative Insulin Pump Guideline.)

Intraoperative glycemic management see Figure 11

Glycemic management in the post-anesthesia care unit see Figure 12

Perioperative use of glucocorticoids. If BG is >180 mg/dL in the preoperative area or patient has significant insulin resistance, the anesthesia team should consider not administering dexamethasone.

Anesthesia and surgery teams should adhere to guidelines developed for administration of supplemental steroids in the perioperative period. At UMHS see Prophylaxis for PONV in the Adult Patient.

Postoperative floor management of diabetes and transition of diabetes care see Figure 12

Patients on an insulin drip in PACU who are moving to an ICU bed should continue on an insulin drip in the ICU.

Patients on an insulin drip in PACU who are moving to a floor bed should be transitioned off the insulin drip as soon as it is safe to do so after arrival on the floor using the following parameters:

-

If no prior history of diabetes (ie, stress hyperglycemia only), insulin drip can be discontinued once it is <2units/hour and the patient should be written for a low or moderate dose correction scale. If correction scale alone does not provide adequate control within 24 hours follow Figure 2 for initiation of scheduled insulin.

-

If the patient has a prior history of Type 2 diabetes but is not on insulin at home (i.e. diet controlled or non-insulin agents only), insulin drip can be discontinued once it is <2units/hour and the patient should be written for a moderate dose correction scale. If correction scale alone does not provide adequate control within 24 hours follow Figure 3 for initiation of scheduled insulin. Patients with A1c >7.5 or on 2 or more oral agents at home may need scheduled insulin initiated sooner.

-

If the patient has a prior history of Type 2 diabetes on basal or basal-bolus therapy at home and the insulin drip is <2units/hour the insulin drip can be discontinued and the patient can be transitioned to SC insulin using Figure 5 or Figure 6. If they are using >2units/hour they should be continued on the insulin drip until their insulin requirements stabilize then transitioned to a basal-bolus regimen. Use Figure 1 to estimate their 24-hour drip requirements taking into account when they last dose their long-acting insulin prior to surgery and the dose, then follow Step 2 and Step 3 to initiate basal-bolus regimen. The insulin drip is discontinued 2 hours after the first injection of long- or intermediate-acting insulin.

Nutritional Issues

Nutrition education

For patients with newly diagnosed diabetes, consult an inpatient Registered Dietitian Nutritionist (RDN) for diabetes nutrition education.

For long term follow up and reinforcement, refer patients to an outpatient diabetes class, RDN, or Certified Diabetes Educator as needed.

Oral nutrition

Give a consistent carbohydrate diet to hospitalized diabetes patients when possible. At UMHS see Table 9.

Glycemic management of patient receiving enteral nutrition. (Table 10)

Total Parenteral Nutrition (TPN)

At UMHS, see Adult UMHS Hospital Guideline for Addition of Insulin to TPN.

Patients with T1DM or T2DM on insulin and TPN may have basal insulin up to 40% of Total Daily Dose (TDD), in addition to TPN insulin.

Other Issues

U-500 Insulin

Do not prescribe U-500 insulin without both approval and consultation from endocrinology. (At UMHS, U-500 insulin cannot be prescribed without approval and consultation from endocrinology).

Hypoglycemia management

Hypoglycemia in the inpatient setting should prompt an evaluation for cause and, if appropriate, a change in existing insulin regimen (Table 2).

At UMHS, manage hypoglycemia in accordance with the Hypoglycemia Treatment Algorithm: Adult

Diabetes education in the hospital setting

Focus on survival skills (Table 3).

Preparing patients for discharge

Modify diabetes therapy at the time of discharge in accord with Figure 4.

Figure 1.

Calculating Insulin Doses and Prescribing

Figure 2.

Clinical Approach to Stress Hyperglycemia with No History of Diabetes

BG = blood glucose; DM = diabetes mellitus; POC = point of care; QAC = before meals; QHS = after meals.

Figure 3.

Management of Type 2 Diabetes on Non-Insulin Agents Only Prior to Admission

BG = blood glucose; DM = diabetes mellitus; eGFR = estimated glomerular filtration rate; POC = point of care; QAC = before meals; QHS = after meals; q = every, NIA- non-insulin anti-diabetic agent

^a. One notable exception is DPP4-inhibitors such as sitagliptin or linagliptin, which can be safely continued on admission if desired.

^b. A1c >7.5 represents glucose control that could be improved. However, in patients with significant comorbidities, A1c <8% may be acceptable.

^c Sitagliptin dose is based on eGFR: ≥ 45 cc/min = 100 mg daily, 30–44 cc/min = 50 mg daily, < 30 cc/min = 25 mg daily.

Figure 4.

Modifying the Diabetes Medication Regimen at the Time of Discharge

NIA = non-insulin antidiabetic agent

Figure 5.

Management of Type 2 Diabetes for Patients on Basal Insulin Only or Basal Insulin Plus Non-Insulin Agents Prior to Admission

^a. One exception is DPP4-inhibitors such as sitagliptin or linagliptin, which can be safely continued on admission if desired.

^b. A1c >7.5 represents glucose control that could be improved. However, in patients with significant comorbidities, A1c <8% may be acceptable.

^c. Basal: use closer to 100% in patients poorly controlled on their home doses, use closer to 75% in patients who report hypoglycemia or who are at risk of hypoglycemia (AKI, malnourished, etc.).

^d. CORRECTION SCALE: Low dose = <40 units, moderate dose = 40-80 units, high dose = >80 units continue correctional insulin scale.

^e. Sitagliptin dose is based on eGFR: ≤45mL/min = 100 mg daily, 30-44mL/min = 50 mg daily, <30mL/min = 25 mg daily.

BG = blood glucose; DM = diabetes mellitus; POC = point of care; QAC = before meals; QHS = after meals; AKI = Acute Kidney Injury, NIA- non-insulin anti-diabetic agent

Figure 6.

Management of Type 2 Diabetes on Basal Bolus Insulin Prior to Admission

^a. A1c >7.5 represents glucose control that could be improved. However, in patients with significant comorbidities, A1c <8% may be acceptable.

^b. BASAL: use closer to 100% in patients on a balanced basal-bolus regimen, use closer to 75% in patients who report hypoglycemia or who are at risk of hypoglycemia (AKI, malnourished, etc.).

^c. IF NPO – Hold bolus insulin. Reduce basal insulin dose to 70% of home dose. Continue correctional insulin scale.

^d. Sitagliptin dose is based on eGFR: ≥45mL/min = 100 mg daily, 30–44mL/min = 50 mg daily, <30mL/min = 25 mg daily.

^e. CORRECTION SCALE: The appropriate correction scale can be determined by considering the patient’s known (or estimated) total daily dose (TDD) of insulin. Low dose = TDD <40units, moderate dose = TDD 40–80 units, high dose = TDD >80 units.

BG = blood glucose; DM = diabetes mellitus; eGFR = estimated glomerular filtration rate; POC = point of care; QAC = before meals; QHS = after meals; AKI = acute kidney injury.

Figure 7.

Glycemic Management in Patients with New Diagnosis of Diabetes

DKA = Diabetic Ketoacidosis

Figure 8.

Consideration of Surgery Postponement Due to Poor Glycemic Control in the Pre-Op Setting

Figure 9.

Day-of-Surgery Considerations for Delay or Cancellation of Procedure

* Small ketones can be due to overnight fasting

** If the patient is in DKA:

• The patient should be admitted to an observation unit (at UMHS, UH South) for DKA management.

• The surgical team is responsible for the management of the patient until they can be moved to the observation unit.

• If the observation unit is full, the surgical team should find an expedited bed (at UMHS speak to ABCC).

Figure 10.

Glucose Management in the Preop/Holding Area

* Perioperative Insulin Pump Guideline for UMHS only

Note: Patients with Type 1 Diabetes must have some basal insulin at all times. If patient did not take reduced dose of long acting insulin per usual preop instructions, consider starting IV insulin infusion in preop or administering reduced dose of patient’s long-acting insulin).

BG = blood glucose

Figure 11.

Intraoperative Glucose Management

* Perioperative Insulin Pump Guideline for UMHS only

** Patients with Type 1 diabetes must have some basal insulin at all times (long-acting Sub-Q or IV infusion)

BG = Blood glucose; Q = every

Figure 12.

Glucose Management in the Post-Anesthesia Care Unit

* Patients with Type 1 diabetes must have some basal insulin at all times (long-acting Sub-Q or IV infusion).

** Perioperative Insulin Pump Guideline for UMHS only

Insulin and Its Components

A wealth of literature and clinical experience exists for using insulin in the hospital setting. Insulin does not interact with other medications and can safely be used in patients with multisystem disease (especially renal and hepatic dysfunction) in appropriately adjusted doses and with modified glycemic targets.

General Principles of Glycemic Management in the Hospital

Insulin is a high-risk medication and, most significantly, can cause hypoglycemia. Following guidelines reduces the risk of adverse side effects and medication errors. Order sets in the electronic medical record can facilitate basal-bolus insulin prescription. Scheduled insulin is to be administered in a physiologic pattern as basal, bolus, and correctional insulin.

1. Basal insulin is given as intermediate or long-acting insulin injected once or twice a day. AM dosing is preferred in patients with end stage renal disease.
2. Bolus insulin is given as rapid or short-acting (RA/SA) insulin, which has 2 components:

   • Meal insulin: RA insulin is used to cover the carbohydrate content of a meal. This can be given 15 minutes before or immediately after finishing a meal (up to 10 minutes). Hospitalized patients have unpredictable oral intake, therefore should preferably receive their RA insulin immediately after eating.
   • Correction insulin: RA insulin is used to lower glucose levels that are above target. Figure 1 shows selection of appropriate correctional scale based on patient’s total daily insulin requirements.

     • If a patient is eating, correction insulin can be given with meal dose in one injection.
     • If patient is not eating and BG is high, correctional insulin should be given alone.
   • For patients on oral agents and/or TDD of insulin of ≤0.5 units/kg/day, sitagliptin or linagliptin can replace bolus meal doses. A correction scale should still be prescribed.

Total daily dose of insulin is calculated based on four options: See Figure 1.

1. Home regimen with modifications: For patients on a home regimen, decrease outpatient doses by 25%, especially if a patient is tightly controlled at home or has risk factors for hypoglycemia (such as poor oral intake or basal heavy regimen).
2. Weight based: Especially for insulin naïve patients, those with high A1c, or with significant hypo- or hyperglycemia on current regimen. Use the appropriate N factor as described in Figure 1.
3. Insulin drip transition: An estimated total daily dose can be calculated to convert IV to subcutaneous insulin therapy.
4. Correction scale: For patients initially on correction scale-only insulin, a total daily dose can be estimated based on their total correction in the last 24 hours.

Patient comorbidities must be considered when calculating insulin doses. ESRD, hypoglycemia unawareness, severe CAD, neurological injury and malnutrition predispose a patient to hypoglycemia.

An appropriate clinic history helps determine the appropriate inpatient glycemic regimen.

Consider the following points:

• Does the patient have a previous diagnosis of diabetes?
• Does the patient have T1DM or T2DM?
• How long has the patient had diabetes?
• What prior agents has the patient been on?
• What is the patient’s current diabetes regimen?
• What is the patient’s adherence to their home regimen?
• Does the patient check his/her BG, if so how often?
• What are the patient’s BGs generally at home?
• What is the patient’s previous A1c?
• Who manages the patient’s diabetes as an outpatient?
• Does the patient have any macrovascular complications (CVD, PVD, stroke)?
• Does the patient have any microvascular complications such as retinopathy, neuropathy, or nephropathy?
• Can interventions in the hospital, including education, supplies, medications, or social interventions, improve DM care at home?

Glycemic monitoring in the hospital.

Patients with or without diabetes can experience significant hyperglycemia in the hospital due to physiologic stress associated with illness. Admission glucose measurement can identify those with stress hyperglycemia or with previously undiagnosed diabetes.^2–5 Also, if BG is elevated, obtain an A1c to differentiate between stress hyperglycemia and newly diagnosed diabetes.

For patients with known diabetes, BG control and medication requirements can vary significantly between the hospital and the home setting. Therefore, frequent monitoring is essential to prevent both hyper- and hypoglycemia.

Perform point-of-care BG checks using the hospital’s glucose meters (not the patient’s home meter or continuous glucose sensor). Document the checks in the medical record. Patients who use continuous glucose monitoring (CGM) devices at home may continue to use these devices in the hospital for self-monitoring. However, this does not replace QAC & QHS point-of-care checks in the hospital. Insulin should always be dosed based on point-of-care readings and never off the CGM device in the hospital.⁶

If BGs are <140 mg/dL for 24 hours with no or minimal insulin correction, BG checks may be discontinued in patients with T2DM on non-insulin agents or stress hyperglycemia. Patients who require insulin (even if their BGs are at target) should continue to have regular monitoring QAC & QHS (or q6 hours if NPO).

Patients started on steroids (≥7.5mg prednisone or equivalent per day), tube feeds, or TPN are at high risk of hyperglycemia and require continued monitoring even if they are euglycemic on admission.^7–10 See the individual sections on steroids, tube feeds, and TPN for further information on these special circumstances.

Hemoglobin A1c.

A1c is a measure of glycated hemoglobin and gives an estimate of glycemic control over the last 3 months. A1c ≥6.5% is consistent with the diagnosis of diabetes, and between 5.7–6.4% suggestive of prediabetes.¹³

A1c levels are useful for patients without and with a diagnosis of diabetes.

• For patients without a previous diagnosis of diabetes and with hyperglycemia in the hospital, an elevated A1c suggests new/unrecognized diabetes rather than stress hyperglycemia.^14–17
• For patients with known diabetes admitted to the hospital, the A1c helps assess prehospitalization glycemic control and, when high, identifies patients needing their medication regimen intensified on discharge or who would benefit from additional instruction on disease management.^18–21

Data are mixed regarding the use of A1c as a predictor of poor outcomes in patients with known diabetes in the inpatient setting.^22–24 For patients with high A1c, the hospitalization may be an opportune time to work on improving glucose control at home.

A1C can be artificially elevated or lowered by several conditions.^25,26

• A1c is elevated by chronic anemia from iron deficiency, B12 or folic acid deficiency, asplenia, uremia, severe hypertriglyceridemia, severe hyperbilirubinemia, chronic alcoholism, chronic aspirin or salicylate use, chronic opioids and lead poisoning.
• A1C is lowered in acute blood loss, hemolytic anemias, splenomegaly, pregnancy, ribavirin, and interferon use.
• Transfusion and hemoglobinopathies have variable effects

Management of Specific Clinical Scenarios

Inpatient Management of Type 2 Diabetes Patients Taking Only Non-insulin Agents Prior to Admission

Discontinue the use of most noninsulin agents during hospitalization except in select stable patients who are expected to have regular meals.²⁸

• Sulfonylureas – discontinue because of the risk of hypoglycemia, as nutritional intake is unpredictable. This is especially true for glyburide which can cause prolonged hypoglycemia.
• Metformin:

  -

  If contrast studies are anticipated, renal function is fluctuating, or estimated glomerular filtration rate (eGFR) < 45, hold on admission.

  -

  If no contraindications are present (eg, renal failure, IV contrast), it may be reinstituted close to discharge.

• Thiazolidinediones (TZDs) – very slow acting and can cause fluid retention and liver toxicity.
• Glucagon-like peptide-1 (GLP-1) agonists – cause nausea and can slow gastric emptying.
• Sodium/glucose cotransporter 2 (SGLT2) inhibitors – cause glycosuria and volume changes.

Sitagliptin and linagliptin, the dipeptidyl peptidase-4 (DPP-4) inhibitor, has been evaluated for use in the inpatient setting. Out of the four DPP-4 inhibitors approved for the outpatient treatment of hyperglycemia alone or in combination with other oral agents, sitagliptin and linagliptin have been evaluated in randomized control trials for hospital use and found to be safe and effective therapy for inpatient management of general medicine and surgery patients with T2DM.^29–32 The Trial Evaluating Cardiovascular Outcomes with Sitagliptin (TECOS), found no adverse cardiovascular outcomes. It needs to be dosed according to the patient’s eGFR.³³

Linagliptin has the advantage of not having any adjustment for renal failure.

In general, reinstitute home oral agents prior to discharge from the hospital if no contraindications develop and BG control was reasonable on them prior to admission (Figure 4). Reduce insulin doses (by about 25%) when oral agents are added back for discharge and BG is well controlled on insulin.

In low acuity patients who are being transferred to inpatient psychiatry, inpatient rehab, or a long-term care facility, consider reinstituting home regimen assuming no contraindications.

Inpatient Management of Type 2 Diabetes Patients Taking Basal-only Insulin, or Basal Insulin Plus Non-Insulin Agents, Prior to Admission

Most patients require about 50% of insulin administered as basal insulin and 50% as bolus insulin. Patients taking high doses of basal-only insulin at home may actually be using their basal insulin to cover some of their meal requirements. Patients on “basal-heavy” regimens are more prone to hypoglycemia when NPO. Switching these patients to a basal-bolus regimen in the hospital is often useful and requires adjustment of basal dose.

Inpatient Management of Type 2 Diabetes Patients Taking Basal-bolus Insulin Prior to Admission

Patients on basal-bolus insulin therapy can generally continue their insulin regimen in the hospital with adjustment as indicated in Figure 6. One notable exception is for patients who are on mixed insulin (see Table 1 for listing of mixed insulins). Mixed insulins contain both intermediate-acting and short-acting insulin combined in a single injection typically dosed twice a day. Patients on this type of insulin are prone to hypoglycemia unless they maintain a consistent meal schedule. Given that PO intake is variable in hospitalized patients it may to be safer to convert to a traditional basal-bolus regimen. To convert, add up the amount of insulin the patient uses per day to get a total daily dose and give 50% as basal insulin and 50% as bolus insulin. For instance, a patient on Novolin 70/30 who takes 50units with breakfast and 40units with dinner could be converted to 45units glargine and 15units lispro TID with meals. A dose reduction of up to 25% from the home total daily dose can still be considered in patients where there is concern for hypoglycemia as indicated in Figure 6.

New Diagnosis of Diabetes (in the hospital)

A new diagnosis of T1DM should be considered in patients presenting with diabetic ketoacidosis (DKA) or patients with hyperglycemia who do not fit the typical pattern for T2DM (patients <35 years or lean patients). Antibodies can be useful to help distinguish T1DM from T2DM. However, if T1DM is suspected, do not delay the initiation of insulin while awaiting the results of antibodies.

GAD-65 antibody tests are a useful first step in patients with suspected T1DM, and other antibody tests can be used on an outpatient basis.³⁴

Endocrine consultation is appropriate for patients with new T1DM because insulin requirements can be highly variable and tend to trend down around the time of discharge (“honeymoon period”). These patients require close follow-up after discharge.

The A1c can help determine which patients require insulin versus which can be managed with lifestyle changes or oral agents. Although the American Diabetes Association suggests a cut off A1c of 10% to start insulin vs oral agents, this may be too high and 9% is more appropriate in the hospital setting.

Inpatient Management of Patients with Type 1 Diabetes

The key difference between T1DM and T2DM patients in the inpatient setting is that basal insulin should NEVER be held in a patient with T1DM as they do not make their own insulin. In the absence of studies specifically addressing inpatient glycemic targets in the T1DM population, for this population adhering to the same BG goals as the T2DM population is reasonable.

The absence of insulin, even for short periods of time (>2 hours), produces a catabolic state and can lead to DKA. DKA can occur at BGs as low as 250 mg/dL. This is especially true in patients both with T1DM and T2DM who are on a SGLT-2 inhibitor (euglycemic DKA). Maintain a high degree of clinical suspicion for DKA in any patient with T1DM or ketosis-prone type 2 who has been without basal insulin for >2–3 hours.

T1DM patients typically count carbohydrates. They can start with their home carbohydrate ratio in the hospital and adjust for high or low BG levels with correction. To help with this calculation, carbohydrate content of meals is on the menu in carbohydrate consistent diets. Insulin for carbohydrates must be prorated to a patient’s actual intake. The patient will need to be able to communicate with nursing regarding their carbohydrate intake and insulin to carbohydrate ratios. The carbohydrate ratio must be documented in medication orders and the amount of insulin given at each meal must be documented in the medication administration record.

Doses for T1DM patients on fixed doses of meal insulin may need to be reduced for decreased appetite or poor oral intake.

If the patient uses a sensitivity factor, this can guide the choice of correction scale. The standard low-dose correction scale at UMHS may be too aggressive for many T1DM patients, especially for those that are insulin sensitive or those with renal insufficiency. In these circumstances, a custom correction scale of 1:50 >150 (ie, 1 unit of insulin for every 50 mg/dL that the BG is greater than 150 mg/dL) may be more appropriate or a custom scale using their home sensitivity factor may be ordered.

Inpatient Management of Patients Using an Insulin Pump (continuous subcutaneous insulin infusion) at Home

An insulin pump is a device that delivers rapid-acting insulin as a slow continuous subcutaneous infusion (basal component). The patient gives boluses of rapid acting insulin via the pump when the patient is eating or has high glucoses (mealtime and correctional component). The patient must prompt the pump to give boluses, either by entering the carbohydrates consumed to get a mealtime bolus or by entering the BG reading to get a correction bolus. Since the pump uses rapid acting insulin, the patient should never suspend or remove their pump for >2 hour (unless long-acting is given in its place).

At UMHS, please see the Insulin Subcutaneous Infusion Pump Therapy Policy for assistance in deciding which patients can appropriately continue their pumps in the hospital, and which patients should be changed to a different management strategy. The admitting service must contact Endocrinology for approval/direction to continue use of the insulin pump therapy during hospitalization. The order set for insulin infusion pump settings and use must be completed in the EHR by the admitting service or Endocrinology. Orders for pump therapy must be placed in the patient records (at UMHS, use MiChart; entry is located as part of the END subcutaneous insulin therapy order set). The pump order set documents the basal rates, carbohydrate ratios, and correction factors used by the pump. It also allows nursing to document BG readings and boluses in the medication administration record.

Patients on a pump must have their BGs checked 4–5 times per day on the hospital BG meters. Checking on a home meter linked to the pump does NOT substitute for checking on the hospital meter, as those readings are not documented in the medical record.

Some insulin pumps are capable of incorporating continuous glucose monitoring data in order to self-adjust insulin dosing. These types of automated insulin delivery systems (including hybrid closed loop technology) has not yet been tested in the hospital. In accordance with the current insulin pump policy at UMHS we recommend that patients on these types of devices have automated insulin delivery turned off and instead use “manual mode” while they are in the hospital. Automated insulin delivery technologies are a rapidly changing field and the endocrine consult service should be involved to help clarify which type of device the patient is using and whether or not it is appropriate for the hospital setting.

Steroid Induced Hyperglycemia

Steroids can cause hyperglycemia in patients with and without DM. Steroids increase mealtime insulin requirements to a greater extent than fasting (basal) insulin requirements.

Options available for controlling hyperglycemia with steroid regimens depend on the steroid doses and degree of hyperglycemia, summarized in Table 5.

Prednisone once daily in the morning causes a BG peak between lunch and dinner. Hyperglycemia in patients on single dose prednisone in the AM can be treated by adding a single dose of NPH timed with steroid administration. If patient is already on basal bolus regimen, NPH can be added as a third insulin in addition to a basal and bolus insulin. This is particularly beneficial for patients on a rapid steroid taper, as the NPH dose can be decreased in a stepwise fashion with each decrease in the prednisone dose without disturbing the basal bolus regimen.

For patients on long acting steroids, or multiple doses of steroids per day, a basal-bolus regimen with a moderate to high dose correction scale may be more useful. The insulin regimen will need to be adjusted as per BG levels and each time steroids are increased or decreased.³⁵

For very high-dose dexamethasone treatments like CVAD (Cyclophosphamide, Vincristine Sulfate, Doxorubicin Hydrochloride [Adriamycin], Dexamethasone), high doses of basal-bolus and correction insulin are necessary with an emphasis on giving adequate meal insulin. An insulin drip may be necessary for uncontrolled hyperglycemia.³⁶

Pregnancy and Peripartum

Perioperative Diabetes Management

Nutrition Education

Oral Nutrition

Table 9 shows consistent carbohydrate diets available at UMHS. If patients request additional food to feel satisfied, protein and fats can be added freely to the diet.

Obstetric patients have different nutritional needs than the typical patient with diabetes. See Table 6 and Table 7 for dietary recommendations based on clinical scenario.

Enteral Nutrition (“Tube Feeds”)

Total Parenteral Nutrition (TPN)

Add regular insulin directly to the TPN solution rather than using subcutaneous insulin to cover TPN. This provides a safety mechanism in that if the TPN is stopped for any reason, the patient will not have excess insulin on board. Adding insulin to the TPN solution also provides maximal flexibility once TPN is cycled to be given over a period of < 24 hours. The ratio of insulin to carbohydrates remains the same, even if the TPN rate changes (ie, the carbs are administered at a faster rate over shorter time and the insulin is also administered at a faster rate over shorter time, so the ratio stays fixed).

T1DM patients are an exception and should ALWAYS have subcutaneous basal insulin at all times in addition to TPN insulin. Basal insulin is also added for patients with T2DM on insulin at home or requiring very high doses of insulin with TPN. [UMHS staff see the UMHS Ambulatory Care Diabetes guideline.]

U-500 insulin

U-500 insulin is five times more concentrated than regular U-100 insulin. It is given 2–3 times per day with meals and takes the place of both basal and bolus insulin. Its use is generally restricted to patients on >200 units of insulin per day. It is highly error prone due to discrepancies in how doses are reported (dosed in milliliters and units).

As part of the history, inquire about what type of syringe the patient uses (TB syringe, U-500 syringe, U-100 syringe, or pen) and how the patient draws up the insulin (eg, “I draw up my insulin using an insulin syringe to the ‘20’ mark” means that the patient is taking 0.2ml or 100 units).

Patients on U-500 often need significant dose reductions in the hospital. They sometimes need to be switched to basal glargine and bolus lispro with meals.

Providers MUST contact endocrinology for consultation prior to use of U-500 in the hospital because of the high risk of medication errors and need for significant dose adjustment in hospital setting. At UMHS, the pharmacy will not dispense without an approving endocrinologist. Staff at UMHS should refer to the Utilization of Insulin Concentrate U-500 Policy for further details.

Hypoglycemia Management

Diabetes Education in the Hospital Setting

Early intervention with ample opportunities for patients to practice skills using teach-back methods reduces readmissions and emergency room visits.⁴⁷

Bedside nurses often have optimal teaching moments when they are administering medications and checking BG levels. The inpatient setting poses a challenge in that patients are acutely ill, under stress and may also be in pain. Inpatient diabetes education should focus on the basic survival skills necessary to safely prepare patients for discharge (Table 3). Keep sessions brief and focused, and minimize interruptions for a more productive learning experience.

Preparing Patients for Discharge

Diabetes and hyperglycemia are commonly not the principle reason for hospitalization. Hospitalization is an opportunity to improve glycemic control in patients with poor control at baseline. Intensification of inadequate regimens during hospitalization can improve long-term glycemic control and reduce hospital readmissions.^1–4

The optimal discharge medication regimen may differ from either the inpatient regimen or the admission regimen. The prehospital regimen is a useful baseline. Inpatients are often treated with insulin, with doses titrated to effect. The inpatient regimen can inform the selection of the discharge regimen as well.

Several factors may affect BG levels after discharge.

• The stress of acute illness or surgery can raise BG. This effect will wane over time as the patient recovers. Either consider this change when developing discharge regimen or arrange for close BG follow-up.
• Significant hyperglycemia can be caused by medications (eg, corticosteroids, thiazide diuretics, calcineurin inhibitors like cyclosporine or tacrolimus, atypical antipsychotics like clozapine or olanzapine). Use of these medications may necessitate adjustment of the diabetes medication regimen.
• Patients with deteriorating renal function can experience prolonged effects of insulin, and might require dose reductions.
• Dietary changes might impact glycemic control after discharge. For example, patients who start on tube feeds often require treatment for glycemic control of tube-feed related hyperglycemia.

Discuss adjustments in the discharge regimen with the patient or caregiver to ensure that they are willing and able to intensify their regimen. The discussion is particularly important for inpatients new to insulin, or who are increasing from a once daily injection to multiple daily injections. All patients need to have basic diabetes knowledge (survival skills) at discharge (Table 3). For new onset DM or new insulin starts, patients will need additional, formal diabetes education after discharge. Patients should be supplied the appropriate supplies at discharge as well (Table 4). For patients with chronic, very poor glycemic control, glycemic targets should be attained gradually to avoid exacerbation of neuropathy or retinopathy.⁴⁸

Provide anticipatory guidance about what to do in the event of high or low BGs after discharge. The guidance can be provided either as titration instructions or as call parameters with contact information for the appropriate physician.

In the discharge summary, providers should document any changes that were made to the diabetes medical regimen, and identify the medical provider who will be responsible for outpatient diabetes management. Close follow-up should be arranged.

Related National Guidelines

The University of Michigan Health System (UMHS) Clinical Guideline on Inpatient Glycemic Management is generally consistent with the guidelines of the American Diabetes Association (ADA). There are some minor differences:

ADA recommends an A1c of 10% as a cutoff for patients diagnosed with new diabetes to transition from lifestyle changes and oral agents to injected agents. For hospitalized patients we recommend a cutoff at A1c of 9% Table 5 due to the expected increased acuity of these patients.

Related National Performance Measures

Relevant national measures for inpatient glycemic management are:

Centers for Medicare and Medicaid Services: Glycemic Control – Hyperglycemia. Average percentage of hyperglycemic hospital days for individuals with a diagnosis of diabetes mellitus, anti-diabetes drugs (except metformin) administered, or at least one elevated glucose level during the hospital stay.

Centers for Medicare and Medicaid Services: Glycemic Control – Hypoglycemia. The rate of hypoglycemic events following the administration of an anti-diabetic agent.

Funding

The development of this guideline was funded by the University of Michigan Health System.

Guideline Development Team and Disclosures

The multidisciplinary guideline development team represented the following units of University of Michigan Health System:

Anesthesiology: Subramanian Sathishkumar, MBBS

General Surgery: Barbra S Miller, MD

Hospital Medicine: David H. Wesorick, MD

Learning Health Sciences (guideline development methodologist): F. Jacob Seagull, PhD

Literature search services: informationists at the Taubman Health Sciences Library

Metabolism, Endocrinology, and Diabetes: Roma Y Gianchandani, MBBS; Jennifer J Iyengar, MD

Nursing Services: Jennifer L Pesenecker, RN

Nutrition Services: Jackie Kosanka, RD

Obstetrics & Gynecology: Mark C Chames, MD

Pharmacy Services: Simona O. Butler, PharmD

University of Michigan Health System endorses the Guidelines of the Association of American Medical Colleges and the Standards of the Accreditation Council for Continuing Medical Education that the individuals who present educational activities disclose significant relationships with commercial companies whose products or services are discussed. Disclosure of a relationship is not intended to suggest bias in the information presented, but is made to provide readers with information that might be of potential importance to their evaluation of the information.

No relevant personal financial relationships with commercial entities: Simona O Butler, PharmD; Mark C Chames, MD; Roma Y Gianchandani, MBBS; Jennifer L Iyengar, MD; Jackie Kosanka, RD; Barbra S Miller, MD; Jennifer L Pesenecker, RN; Subramanian Sathishkumar, MBBS; F. Jacob Seagull, PhD; David Wesorick, MD.

Relevant personal financial relationships with commercial entities: None.

Strategy for Literature Search

Two different Main searches were used to search this topic, and the strategies are contained in the Search Strategies document. The second strategy was used in searching type 1 diabetes, and stress hyperglycemia. The comprehensive main search on all glycemic management in diabetes and hyperglycemia retrieved 2,486 references. When the search hedges for Guidelines, Clinical Trials, and Cohort Studies were added, the base results are as follows:

Glycemic Management -Guidelines, total results were 111

Glycemic Management -Clinical Trials, total results were 403

Glycemic Management -Cohort Studies, total results were 834

A search on all glycemic management articles that weren’t covered in the rest of the searches was added. The search retrieved 408 articles.

When the search hedges for Guidelines, Clinical Trials, and Cohort Studies were added, the base results are as follow:

Other Glycemic Management -Guidelines, total results were 33

Other Glycemic Management -Clinical Trials, total results were 72

Other Glycemic Management -Cohort Studies, total results were 107

The MEDLINE In-Process database was also searched using the strategy in the search strategies document. The search retrieved 154 documents. The results with the hedges applied are:

Guidelines, total results were 1

Clinical Trials, total results were 11

Cohort Studies, total results were 43

Within the Cochrane Database of Systematic Reviews, 64 reviews were found using the strategy in the search strategies document. Many of the Cochrane results are not relevant, but were retrieved using a strategy similar to what was used in Medline.

Level of evidence supporting a diagnostic method or an intervention:

A = systematic reviews of randomized controlled trials

B = randomized controlled trials

C =systematic review of non-randomized controlled trials or observational studies, non-randomized controlled trials, group observation studies (eg, cohort, cross-sectional, case control)

D = individual observation studies (case or case series)

E = opinion of expert panel.

Search details and evidence tables available at http://www.uofmhealth.org/provider/clinical-care-guidelines.

Recommendations

Guideline recommendations were based on prospective randomized controlled trials if available, to the exclusion of other data; if RCTs were not available, observational studies were admitted to consideration. If no such data were available for a given link in the problem formulation, expert opinion was used to estimate effect size. The “strength of recommendation” for key aspects of care was determined by expert opinion.

The strength of recommendations regarding care were categorized as:

I = Generally should be performed

II = May be reasonable to perform

III = Generally should not be performed

Data Availability

Internal UMHS Guidelines contain proprietary information so are only accessible to appropriate Michigan Medicine staff. For more information, contact the authors or publisher.

Review and Endorsement

Unit review: Drafts of this guideline were reviewed in clinical conferences and by distribution for comment within departments and divisions of the University of Michigan Health System to which the content is most relevant: Anesthesiology; Family Medicine; General Surgery; Hospital Medicine; Metabolism, Endocrinology, and Diabetes; Nursing Services; Nutrition Services; Obstetrics & Gynecology; and Pharmacy Services.

Institutional review: The final document was distributed for institutional review by the Clinical Practice Committee of the University of Michigan Medical Group and by the Executive Committee for Clinical Affairs of the University of Michigan Hospitals. These committees represent a broad spectrum of medical specialties and other health care professionals.