In the hospital setting, hyperglycemia is often seen in patients with and without a diabetes diagnosis before admission. Current guidelines call for aggressive treatment of inpatient hyperglycemia, and continuous intravenous insulin is the therapy of choice, specifically when glycemic control must be reached quickly. Once the patient is stabilized, subcutaneous insulin therapy can be administered to approximate physiologic insulin needs. No standardized clinical protocol is available for inpatient subcutaneous insulin administration. Although discredited in the literature, the use of sliding-scale insulin regimens is common, even though such regimens do not meet physiologic insulin needs, are ineffective at best, and dangerous at worst. This approach to the management of hyperglycemia reflects practice patterns before the current understanding of physiologic insulin replacement. Education and support of hospital staff in the use of physiologic basal and nutritional/prandial insulin replacement regimens result in better glycemic control and improved patient outcomes.

In-hospital hyperglycemia negatively affects patient outcomes and healthcare costs, particularly in critically ill patients. Inpatient hyperglycemia is a significant problem, regardless of one’s diabetes history, and can affect morbidity and mortality if not managed appropriately. In a series of 2030 consecutive adults admitted to a hospital in Atlanta during a 15-week period, 1886 patients had blood glucose measurements taken during their stay. Among these, 38% had hyperglycemia (fasting blood glucose ≥126 mg/dL or random blood glucose ≥200 mg/dL on 2 occasions). Only 26% of the patients had a previously confirmed diagnosis of diabetes, and 12% had newly identified hyperglycemia.¹ In-hospital mortality was significantly higher in the patients with newly identified hyperglycemia than in those with a history of either diabetes or euglycemia (P <.01 for both). Furthermore, the patients with newly identified hyperglycemia had longer hospital stays, were more likely to be cared for in an intensive care unit (ICU), and more frequently required continued care in a nursing home after discharge compared with other patients.¹

Similar results have been observed in other studies—hyperglycemia at admission has been found to be associated with increased mortality after myocardial infarction (MI).^2,3 Increased morbidity and overall mortality, increased hospital costs poststroke,⁴ and an increased risk of nosocomial infection following major cardiovascular or abdominal surgery are also associated with hyperglycemia at admission.⁵

Inpatient Glycemic Control Guidelines
Insulin is the most effective agent for the management of hyperglycemia and is routinely administered to hospitalized patients with or without a previous diagnosis of diabetes.

At a 2003 consensus conference sponsored by the American Association of Clinical Endocrinologists (AACE) and the American Diabetes Association (ADA), the American College of Endocrinology (ACE) Task Force on Inpatient Diabetes and Metabolic Control established the following strict guidelines for glycemic control in hyperglycemic hospitalized patients.⁶

For nonpregnant patients in the ICU, the upper-limit goal for blood glucose levels is 110 mg/dL; for other inpatients, the upper-limit target for preprandial glucose is 110 mg/dL, and maximal glucose levels should not exceed 180 mg/dL.⁶ The recommended upper limits for glycemic targets in pregnant patients are preprandial, 100 mg/dL; 1-hour postprandial, 120 mg/dL; and during labor and delivery, 100 mg/dL.

Insulin is considered the most effective agent for maintaining glycemic control in hospitalized patients, regardless of preadmission antidiabetic therapy. Achieving glycemic control with insulin reduces hospital morbidity and mortality after open heart surgery,⁷ coronary artery bypass grafting,⁸ and acute MI.⁹ The ACE guidelines recommend continuous insulin infusion therapy in patients whose glucose level must be brought under control promptly, including critically ill patients and those on prolonged nothing-by-mouth (NPO) status (Table 1).⁶ Subcutaneous insulin therapy should be used to manage hyperglycemia in stabilized patients on general medical floors.¹⁰

Glycemic Goals Often Not Reached
In a recent study, 90 consecutive hospitalized adult patients treated with sliding-scale insulin (SSI) were evaluated for their glucose outcomes.¹¹ SSI regimens were never adjusted in 73 of the patients (81%). During 5 days of therapy, the proportion of patients who attained targeted glycemic control (90-130 mg/dL) ranged from 2% to 10% (mean, 6%). In general, insulin doses remained subtherapeutic, reflected by glucose levels that remained elevated after 523 sliding-scale injections (84% of the insulin shots). Furthermore, there were 10 episodes of hypoglycemia associated with the SSI therapy.

In a separate study, 656 patients who had an acute ischemic stroke were admitted to an urban hospital during a 5-year period. Hyperglycemia (defined as random serum glucose >130 mg/dL) was seen in 40% of the patients at admission.⁴ The majority (90%) continued to have hyperglycemia while in the hospital (mean blood glucose, 206 mg/dL). This may reflect the lack of antidiabetic medications used during hospitalization in 43% of patients (mean serum glucose, 163 mg/dL), treatment with only oral antidiabetic drugs in 10% of the patients (mean serum glucose, 222 mg/dL), and use of only SSI in 13% (mean serum glucose, 208 mg/dL). Thus, although effective treatment is available and recommendations for basal and nutritional insulin may be readily obtained, antidiabetic medications appear to be underused or administered inappropriately in the hospital setting.

Management Strategies
Continuous insulin infusion therapy
The appropriate continuous insulin infusion protocol for a given hospital depends on many factors, including the patient population and staff resources. Specifically, the type of diabetes involved, the coordination of treatment with nutritional therapy, the feeding status of the patient (ie, NPO, total parenteral nutrition, overnight cycling of enteral feedings), degree of insulin resistance, and risk of hypoglycemia must all be considered.¹² Table 2 presents an appropriate insulin infusion protocol.

Subcutaneous insulin regimens
Subcutaneous insulin regimens are used for hyperglycemic inpatients who do not meet the criteria for continuous insulin infusion therapy.⁶ Ideally, subcutaneus insulin regimens should approximate physiologic insulin activity by meeting basal and nutritional/prandial insulin needs.⁶ Basal insulin replaces the low, steady amount of endogenous insulin that is needed to prevent intermeal and nocturnal gluconeogenesis and ketogenesis. Nutritional insulin replacement provides coverage for hyperglycemia associated with intravenous (IV) dextrose, total parenteral nutrition, enteral feedings, nutritional supplements, and discrete meals. When discrete meals comprise the patient’s total nutritional intake, the nutritional and prandial insulin requirements are the same.¹⁰ Critically ill patients often require higher than normal insulin doses because of illness-related insulin resistance and the use of medications that elevate blood glucose levels.^6,10

Currently, no standardized guidelines are available for administration of subcutaneous insulin therapy for stabilized patients on general medical floors, and many institutions rely on SSI dosage regimens even though they have been discredited in the literature for the past 40 years.^13-16

Sliding-scale insulin therapy
Despite increasing questions about its clinical benefit, SSI therapy remains a common approach for glucose control.¹³ Typically, SSI regimens consist of 4 to 6 injections of regular human insulin (Humulin R, Novolin R) daily, with the dose based on blood glucose measurements that are taken 4 times daily before meals and at bedtime. Alternatively, many clinicians use this type of regimen for NPO patients. In both scenarios, the insulin provided is used for basal and prandial insulin replacement, is injected based on retrospective hyperglycemia, and causes “insulin stacking”—administering insulin doses too close in time, before full absorption and before the peak activity of the previous dose is reached—because of the pharmacodynamics of regular human insulin.¹⁷

No standards are available for SSI regimens because, despite their widespread use, they are not considered as acceptable insulin strategies due to lack of evidence of efficacy. Indeed, most clinicians provide insulin doses based on previous experience, which is often related to inappropriate use during training. Many physicians usually use 1 or perhaps 2 different types of SSI orders. However, these methods do not take into account intra- or interpatient glycemic variability.

When metabolic dysregulation in critical illness and/or diabetes results in hyperglycemia, SSI is often implemented with the misperception that this regimen can achieve glycemic control while avoiding hypoglycemia. In reality, however, use of SSI has been shown to create a greater risk of peaks and valleys in blood glucose, with little ability to promote glycemic control.^11,13,18,19 Inpatient insulin therapy ideally should stabilize blood glucose levels, accounting for expected nutritional intake, but SSI aims to correct existing hyperglycemia in a retrospective manner.

SSI regimens use short-acting regular human insulin and do not have a true basal insulin component. These regimens are reactive and allow correction of the insulin dose only after hyperglycemia has already occurred. Thus, these regimens fail to proactively provide coverage for nutritional exposure and diurnal variations in insulin requirements, or rapid changes in illness-related insulin requirements.^10,18 This nonphysiologic approach places patients at risk of large fluctuations in blood glucose levels that may exacerbate hyperglycemia and increase the risk of hypoglycemia when reductions in nutritional intake occur or when insulin injections are stacked.^12,17,18 Another potentially detrimental factor is the disregard for current standards of glycemic control: most SSI regimens call for insulin administration at blood glucose levels of >200 mg/dL, which is well above the recommended guidelines. Indeed, with most SSI regimens, the treating physician is alerted only if the patient’s blood glucose falls below 60 mg/dL or rises above 400 mg/dL.¹⁶

Especially when used without scheduled insulin for meals, SSI results in more significant glycemic variability. Accumulating data show that this type of “rollercoaster” approach to diabetes leads to a greater accumulation of oxidative stress, the same mechanism that is now thought to fuel diabetic vasculopathy.¹⁰ And although insulin appears to have strong antiinflammatory properties, pharmacologic hyperinsulinemia, when present with hyperglycemia, leads to a proinflammatory state, whereas euglycemia and insulin have the opposite effect.²⁰ This may explain why positive patient outcomes were demonstrated in studies in which patients’ glucose levels were well controlled with insulin,²¹ and studies in which insulin was not optimally provided to control glycemia demonstrated negative outcomes.²²

Significant evidence suggests that for prevention of certain complications, there is a critically important but brief window of time wherein glycemic control is crucial to clinical outcomes. To ensure control within this window, the treatment of choice is IV insulin infusion. Observational data suggest that limiting or preventing hyperglycemia in many clinical situations in the early days after the intervention effectively produces better outcomes many days later. These situations include deep sternal wound infection and mortality in cardiac surgery,^7,21 MI,⁹ nosocomial infection rate in surgical cases,⁵ as well as other types of infections linked to postoperative situations or trauma.

To address the limitations of SSI and the importance of euglycemia in hospitalized patients, a comprehensive MEDLINE search was done using the terms “SSI,” “sliding scale,” and “sliding” combined with “insulin.”¹⁶ The majority of the 52 studies identified that were published between 1966 and 2003 indicated that SSI does not effectively control blood glucose in patients with diabetes.¹⁶ An updated search through 2005 also failed to find any clinical studies showing a benefit with SSI regimens.

The most comprehensive clinical study describing the lack of efficacy of SSI regimens was a prospective cohort study, whose objective was to identify predictors of hypoglycemic and hyperglycemic episodes, with special attention given to SSI regimens.¹³ The study included 171 adults with diabetes identified as a comorbid condition during hospital admission over a 7-week period. During their stay, 23% of the patients had hypoglycemic and 40% had hyperglycemic episodes. Patients placed on an SSI regimen had a 3-fold increased risk of hyperglycemia compared with those who did not receive glycemic control therapy. Despite this difference, most patients continued taking their initial SSI regimen for the duration of hospitalization.¹³ Another study showed that adding SSI to routine diabetes medications (ie, oral antidiabetic drugs or a standing dosage of intermediate-acting insulin and/or regular human insulin) in 153 patients with diabetes hospitalized for other conditions did not prevent hyperglycemia or hypoglycemia.¹⁸ And, at a tertiary healthcare center, 62% of hyperglycemic events were the result of improper disease management.¹⁹ Of these events, 26% were due to the exclusive use of an SSI regimen, 35% due to the failure of the physician to accept responsibility for managing the patient’s diabetes while in the hospital, and 15% due to the physician’s own perceived lack of expertise in managing diabetes.

Scheduled insulin therapy
Scheduled, or programmed, insulin regimens provide coverage for basal and nutritional/prandial insulin requirements by using a combination of intermediate- or long-acting insulin or insulin analogs with either short-acting regular human insulin or rapid-acting insulin analogs (Table 3) to approximate the physiologic patterns of endogenous insulin.¹⁰ Supplemental or correction insulin doses can be added to the basal-nutritional/prandial insulin regimen to control premeal hyperglycemia.¹²

Strategies for managing inpatient hyperglycemia can be found in the technical review published in conjunction with the AACE/ADA consensus conference.¹⁰ In general, however, for patients who are eating discrete meals, basal insulin can be replaced using a long-acting insulin, such as insulin glargine (Lantus) or insulin detemir (Levemir), both with long duration of action (approximately 24 hours for insulin glargine and slightly less for insulin detemir) and no pronounced peak. Rapid-acting insulin analogs, such as insulin lispro (Humalog), insulin aspart (NovoLog), or insulin glulisine (Apidra), should be administered just before meals to provide prandial coverage.¹⁰ If quantity of food intake cannot be predicted, it is reasonable for a rapid-acting insulin analog to be administered promptly after the meal. This is not ideal, however, since the insulin does not take effect immediately. Alternatively, a combination of intermediate-acting neutral protamine Hagedorn (NPH) and short-acting human insulin (Humulin 70/30, Novolin 70/30) can be used for basal-prandial coverage. However, NPH (Humulin N, Novolin N) and regular human insulin have peaks of activity that can result in higher insulin action when they are combined, thus resulting in greater rates of hypoglycemia.¹⁰

Rapid-acting insulin analogs are poorly suited for prandial coverage in patients with NPO status, or those receiving total parenteral nutrition, enteral feeds, or overnight cycling of enteral feeds. In addition, there may be a risk in using a long-acting analog, such as insulin glargine, if carbohydrate exposure is suddenly decreased or discontinued. If insulin glargine is continued during total parenteral nutrition or enteral feeds, the dose may need to be smaller than the usual dose-determining rule of 50% of the total daily dose for basal insulin. A dose based on units/kg of body weight is safer for determining the true basal dose when nutritional support is being given. The use of NPH or mixtures (premixes of NPH with regular insulin or rapid-acting analogs), given in small doses and timed every 6 hours, may provide a better margin of safety, because each dose is smaller and shorter acting.¹⁰

Can prescribing habits be changed?
Although the use of SSI regimens is established in many institutions, physicians can be discouraged from using this inappropriate method of managing inpatient hyperglycemia if they are given more effective and convenient options. In one hospital, fewer SSI regimens were ordered after an endocrinology consult service endorsed a minimal-intervention supplemental insulin regimen.¹⁴ This preformatted option was added to an electronic order screen as part of a computerized medical record system at a Veterans Affairs hospital. This supplemental insulin regimen was designed to be used in conjunction with the patient’s preadmission diabetes medications. Before the introduction of this regimen, 97% of the orders written for insulin were for SSI regimens compared with 64% written following the implementation of the new minimal-intervention protocol. No data were collected, however, concerning the impact of this change on glycemic control and patient outcomes.

A significant improvement in glycemic control (P <.01 versus historic controls) was seen in hyperglycemic inpatients at a medical center after a basal-prandial approach to insulin management was implemented to replace the use of SSI.¹⁵ After training in the basal-prandial insulin protocol, residents were given responsibility for the management of inpatients with hyperglycemia. According to the protocol, patients with blood glucose levels of <200 mg/dL were maintained on preadmission diabetes medications (insulin or oral antidiabetic drugs), with some modifications to the oral antidiabetic drugs as medically necessary (ie, impaired renal and liver function, NPO status). The oral drugs were discontinued, and the patient switched to insulin when blood glucose was >200 mg/dL. All insulin-treated patients were started on a basal-prandial regimen, using either NPH and regular human insulin twice daily or insulin glargine and insulin lispro or insulin aspart.¹⁵ Compared with historic controls, patients treated with the new protocol achieved better glucose control, had a lower incidence of blood glucose levels of <250 mg/dL, and had a greater incidence of blood glucose levels of >60 mg/dL. The proportion of hypoglycemic episodes that required IV dextrose was similar in the 2 groups. None of the patients exhibited clinically severe hypoglycemia, and all episodes resolved satisfactorily.¹⁵

Establishing clear and consistent protocols for inpatient hyperglycemia management while minimizing hypoglycemia is likely to increase the ease and comfort with which the hospital staff provides or augments scheduled insulin. As the new protocols are implemented, educating and empowering the nursing and pharmacy staffs may increase staffing time and may generate some reluctance to change. However, interdisciplinary support as it relates to inpatient hyperglycemia management is an important part of implementation of successful treatment protocols.

Conclusion
Glycemic control guidelines call for aggressive management of inpatient hyperglycemia, with continuous insulin infusion as the therapy of choice for situations in which glycemic control must be obtained quickly. Subcutaneous insulin therapy, which should attempt to approximate physiologic insulin activity, can be used once the patient is stabilized on a general medical floor. SSI therapy is frequently used in the hospital setting but is unsafe, provides no benefit, and may induce harm. Additional studies are needed to address these issues and determine the target blood glucose threshold to optimize outcomes in hospitalized patients.

Disclosure statement
Dr Hirsch is a consultant for Abbott Diabetes Care, Eli-Lilly, and Roche and receives grant support from Medtronic Diabetes and Sanofi-Aventis. Dr Braithwaite is a consultant for Novo-Nordisk, is on the Speakers’ Bureau of Sanofi-Aventis, and receives grant support from Bristol-Myers Squibb.

Self-assessment test
1. Which of the following levels is the upper limit for target preprandial glucose for patients in noncritical care units?
A. 100 mg/dL
B. 110 mg/dL
C. 120 mg/dL
D. 180 mg/dL

2. Continuous insulin infusion therapy is indicated for control of hyperglycemia in all the following patient populations, except:
A. Stabilized patients on general medical floors
B. Critically ill patients
C. Patients on prolonged NPO status
D. Women in labor

3. Which of the following insulin types is used in SSI regimens?
A. Long-acting
B. Intermediate-acting
C. Short-acting
D. Rapid-acting combined with insulin analog

4. All the following statements about SSI regimens are true, except:
A. These methods do not take into account intra- or interpatient glycemic variability
B. These methods have a low risk for hypoglycemia
C. SSI aims to correct existing hyperglycemia in a retrospective manner
D. These regimens fail to provide coverage for nutritional exposure and diurnal variations

5. In scheduled therapy, which of the following insulin preparations can be used to replace basal insulin levels?
A. Insulin glulisine
B. Insulin lispro
C. Insulin aspart
D. Insulin glargine

(Answers at end of references list)

References
1. Umpierrez GE, Isaacs SD, Bazargan N, et al. Hyperglycemia: an independent marker of in-hospital mortality in patients with undiagnosed diabetes. J Clin Endocrinol Metab. 2002; 87: 978-982.

2. Bolk J, van der Ploeg T, Cornel JH, et al. Impaired glucose metabolism predicts mortality after a myocardial infarction. Int J Cardiol. 2001; 79:207-214.

3. Capes SE, Hunt D, Malmberg K, et al. Stress hyperglycaemia and increased risk of death after myocardial infarction in patients with and without diabetes: a systematic overview. Lancet. 2000; 355:773-778.

4. Williams LS, Rotich J, Qi R, et al. Effects of admission hyperglycemia on mortality and costs in acute ischemic stroke. Neurology. 2002; 59: 67-71.

5. Pomposelli JJ, Baxter JK III, Babineau TJ, et al. Early postoperative glucose control predicts nosocomial infection rate in diabetic patients. JPEN J Parenter Enteral Nutr. 1998; 22: 77-81.

6. Garber AJ, Moghissi ES, Bransome ED Jr, et al, for the American College of Endocrinology Task Force on Inpatient Diabetes Metabolic Control. American College of Endocrinology position statement on inpatient diabetes and metabolic control. Endocr Pract. 2004;10(suppl 2):4-9.

7. Furnary AP, Wu Y, Bookin SO. Effect of hyperglycemia and continuous intravenous insulin infusions on outcomes of cardiac surgical procedures: the Portland Diabetic Project. Endocr Pract. 2004; 10(suppl 2): 21-33.

8. Golden SH, Peart-Vigilance C, Kao WH, et al. Perioperative glyce­mic control and the risk of infectious complications in a cohort of adults with diabetes. Diabetes Care. 1999; 22: 1408-1414.

9. Malmberg K. Prospective randomised study of intensive insulin treatment on long term survival after acute myocardial infarction in patients with diabetes mellitus. DIGAMI (Diabetes Mellitus, Insulin Glucose Infusion in Acute Myocardial Infarction) Study Group. BMJ. 1997; 314: 1512-1515.

10. Clement S, Braithwaite SS, Magee MF, et al, for the American Diabetes Association Diabetes in Hospitals Writing Committee. Management of diabetes and hyperglycemia in hospitals [published corrections appear in Diabetes Care. 2004;27:856 and in Diabetes Care. 2004;27: 1255]. Diabetes Care. 2004; 27:553-591.

11. Golightly LK, Jones MA, Hamamura DH, et al. Management of diabetes mellitus in hospitalized patients: efficiency and effectiveness of sliding-scale insulin therapy. Pharmacotherapy. 2006; 26: 1421-1432.

12. Trence DL, Kelly JL, Hirsch IB. The rationale and management of hyperglycemia for in-patients with cardiovascular disease: time for change. J Clin Endocrinol Metab. 2003; 88:2430-2437.

13. Queale WS, Seidler AJ, Brancati FL. Glycemic control and sliding scale insulin use in medical inpatients with diabetes mellitus. Arch Intern Med. 1997; 157:545-552.

14. Achtmeyer CE, Payne TH, Anawalt BD. Computer order entry system decreased use of sliding scale insulin regimens. Methods Inf Med. 2002; 41:277-281.

15. Baldwin D, Villanueva G, McNutt R, et al. Eliminating inpatient sliding-scale insulin: a reeducation project with medical house staff. Diabetes Care. 2005; 28:1008-1011.

16. Browning LA, Dumo P. Sliding-scale insulin: an antiquated approach to glycemic control in hospitalized patients. Am J Health Syst Pharm. 2004; 61:1611-1614.

17. Hirsch IB. An endocrinologist’s view on the practical use of insulin. Insulin. 2006; 1(suppl A): 518-523.

18. Dickerson LM, Ye X, Sack JL, et al. Glycemic control in medical inpatients with type 2 diabetes mellitus receiving sliding scale insulin regimens versus routine diabetes medications: a multicenter randomized controlled trial. Ann Fam Med. 2003; 1:29-35.

19. Smith WD, Winterstein AG, Johns T, et al. Causes of hyperglycemia and hypoglycemia in adult inpatients. Am J Health Syst Pharm. 2005; 62:714-719.

20. Hirsch IB. Inpatient diabetes: review of data from the cardiac care unit. Endocr Pract. 2006; 12(suppl 3):27-34.

21. Furnary AP, Gao G, Grunkemeier GL, et al. Continuous insulin infusion reduces mortality in patients with diabetes undergoing coronary artery bypass grafting. J Thorac Cardiovasc Surg. 2003; 125:1007-1021.

22. Mehta SR, Yusuf S, Diaz R, et al, for the CREATE-ECLA Trial Group Investigators. Effect of glucose-insulin-potassium infusion on mortality in patients with acute ST-segment elevation myocardial infarction: the CREATE-ECLA randomized controlled trial. JAMA. 2005; 293:437-446.

Answers: 1. B; 2. A; 3. C; 4. B; 5. D.