Management of Hyperglycemia in the Noncritical Care Setting: A Real-World Case-Based Approach With Alternative Insulin- and Noninsulin-Based Strategies

A 64-year-old man with a history of hypertension, coronary artery disease, cerebrovascular accident, chronic kidney disease, and alcohol-induced cirrhosis with complications of ascites, jaundice, esophageal varices, and hepatic encephalopathy presented for a liver transplant. He did not have a history of diabetes before the liver transplant and had an A1C of 4.0% at presentation in the setting of hemoglobin of 7.5 mg/dL.

He was started on high-dose intravenous (IV) corticosteroids perioperatively after the liver transplant and developed immediate postoperative hyperglycemia requiring IV insulin infusion. He remained on the IV insulin infusion per the intensive care unit protocol (target glucose of 110–140 mg/dL) for 5 days, initially up to 60 units/day and decreasing to 30 units/day as his steroid regimen was tapered.

On day 5, when an oral diet was initiated and the steroid dose was tapered to prednisone 20 mg daily, he was transitioned from the IV insulin infusion to a BB subcutaneous insulin regimen. He was started on NPH insulin 12 units (dose calculated based on weight) at the time of each steroid administration, along with prandial rapid-acting insulin 4 units (dose calculated based on weight and NPH insulin dose) before each meal, with an additional correction scale when needed.

The patient’s glycemic outcomes remained variable with episodes of hypoglycemia from poor and unpredictable oral intake. This situation led to frequent adjustments of both NPH and prandial insulin doses. Eventually, the prandial insulin was discontinued when the prednisone dose was tapered to 10 mg/day, but postprandial hyperglycemia persisted.

At this time, he was started on linagliptin, a dipeptidyl peptidase 4 (DPP-4) inhibitor, 5 mg daily in addition to NPH insulin 6 units daily in the morning with the prednisone dose. This combination regimen significantly reduced his glycemic variability compared with the previous BB regimen (Figure 1). Subsequently, favorable glycemic outcomes were achieved, with a glucose target of 110–180 mg/dL, no daytime hyperglycemia, and no further requirement for prandial insulin. The patient was eventually discharged on the same regimen for his steroid-induced hyperglycemia.

A 64-year-old woman with a history of type 2 diabetes for 6 years, hypertension, stroke, right below-knee amputation, and peripheral arterial disease with left foot gangrene and revascularization of the left lower extremity a month earlier presented to the emergency department with acute-onset aphasia and right-side weakness and was found to have had a left middle cerebral artery stroke. She did not receive tissue plasminogen activator because of a lack of clarity with regard to the onset of her symptoms, but she eventually underwent endovascular thrombectomy with successful revascularization. Her outpatient diabetes medications included glipizide and metformin. Her current laboratory test values included an A1C of 9.0% (reference range 4.0–5.6%) and a glucose value of 292 mg/dL (65–99 mg/dL). Additional presenting laboratory findings were notable for anion gap metabolic acidosis with a gap of 24 mEq/L (7–15 mEq/L), carbon dioxide level of 14 mEq/L (24–31 mEq/L), lactic acid level of 3.9 mmol/L (0.5–2.2 mmol/L), β hydroxybutyrate level of 4.34 mmol/L (0.02–0.27 mmol/L), estimated glomerular filtration rate of 53 mL/min/1.73 m² (≥90 mL/min/1.73 m²).

For her initial hyperglycemia management, she received a bolus dose of regular insulin along with an IV insulin infusion, with prompt improvement. Subsequently, she was transitioned to once-daily basal insulin glargine 8 units along with low-dose lispro per a correction scale while taking nothing by mouth. Upon initiation of enteral tube feedings, she again developed hyperglycemia that prompted initiation of NPH insulin 6 units (weight-based calculation) every 8 hours along with lispro per the correction scale (weight-based calculation) every 4 hours. As enteral feedings were advanced to goal, she was noted to have progressive hyperglycemia, and her NPH insulin doses were titrated up accordingly (Figure 2). Eventually, she was transitioned to bolus feedings five times per day at 5-hour intervals. This change required a further change in her insulin regimen to NPH insulin twice daily and bolus rapid-acting lispro insulin dosed five times per day before each bolus feeding (Figure 3).

The patient had a prolonged, complicated hospital course with continuing encephalopathy from the stroke. A speech and swallow evaluation was notable for oropharyngeal dysphagia with poor bolus acceptance, reduced labial seal, reduced bolus manipulation, and delayed transit of the bolus, and oral stasis was also noted with puree consistency. Given these findings, it was not considered safe to initiate an oral diet. She eventually underwent a percutaneous gastrostomy tube placement and was discharged to home hospice with bolus feedings five times per day.

Given the patient’s largely acceptable glycemic control while on her inpatient regimen, she was discharged to home hospice on twice-daily intermediate-acting basal NPH insulin and rapid-acting insulin dosed before each bolus enteral feeding five times per day. This case provides an example of an unconventional BB insulin regimen in which bolus insulin was administered up to five times per day with doses given 4–5 hours apart before each bolus enteral feeding to address postprandial hyperglycemia. This regimen achieved the desired glycemic goal of 110–180 mg/dL without hypoglycemia.

A 69-year-old man was admitted with progressively worsening shortness of breath, weakness, and edema. He had a history of type 2 diabetes for 29 years that was complicated by nephropathy leading to dialysis-requiring end-stage renal disease, retinopathy, and neuropathy. His medical history was also pertinent for coronary artery disease with ischemic cardiomyopathy, hypertension, liver cirrhosis, hypothyroidism, and paroxysmal atrial fibrillation.

He had been on peritoneal dialysis (PD) at home and was using a hybrid closed-loop insulin delivery system combining an insulin pump with continuous glucose monitoring. His total daily dose (TDD) of insulin was 65–70 units, with 55% delivered as basal insulin. He used more frequent boluses during his PD hours and was maintained on one basal rate (Figure 4). His most recent A1C was 5.8%.

During hospitalization, his shortness of breath and fluid overload worsened, so hemodialysis (HD) was initiated. The insulin pump was discontinued, and he was placed on a BB insulin regimen with a 30% reduction in his TDD. He only required 2 units for the next 24 hours and then no insulin during the next day, with stable glucose outcomes (Figure 5). His glucose management remained stable while off of insulin. Throughout the rest of the hospital course, his insulin requirements remained <25% of his usual requirements while on PD. He required a very small amount of insulin based on a correction scale.

The reduction in his insulin requirements likely was because of the significant reduction in his dextrose load resulting from the change from PD to HD, as well as his poor oral intake. He was discharged home on an HD regimen, and his insulin pump was reprogrammed at the newly reduced TDD requirements to account for the new, significant changes in his dialysis.

Hyperglycemia is known to be an independent risk factor for both mortality and morbidity in the inpatient setting. Although insulin therapy is the recommended treatment for critically and noncritically ill inpatients, it is also associated with an increased risk of hypoglycemia. Recently, there has been increased interest in alternative glucose-lowering strategies to achieve glycemic outcomes with lower risks of hypoglycemia.

Incretin-based therapies such as glucagon-like peptide 1 (GLP-1) receptor agonists and DPP-4 inhibitors enhance secretion of insulin by restraining the activity of DPP-4 in a glucose-dependent manner and thus lowering the risk of hypoglycemia in hospitalized patients (10–13). DPP-4 inhibitors and GLP-1 receptor agonists have been shown to effectively reduce mean postprandial glucose levels among patients with steroid-induced hyperglycemia without a risk of hypoglycemia, decrease excessive glycemic excursions, and reduce variability in glucose outcomes, thereby reducing or even eliminating the need for insulin (12–14). For transplant patients, both calcineurin inhibitors and corticosteroids are commonly used. Studies have shown that these immunosuppressive agents directly inhibit β-cell insulin secretion by causing β-cell apoptosis, resulting in hyperglycemia and post-transplantation diabetes. Incretin therapies directly counteract the adverse effects of these medications by enhancing insulin secretion by the β-cell and preserving β-cell function. Decreased glucagon secretion and improved peripheral insulin resistance are other beneficial effects of incretin-based treatments that also benefit people with steroid-induced hyperglycemia and confer a potent glucose-lowering effect (15). The evidence for the clinical utility of these mechanisms of action suggests a need for a more detailed investigation on the benefits of incretin-based treatments as alternatives or additions to multiple daily injections of insulin in hospitalized patients (15).

Another common inpatient hyperglycemia scenario includes the management of hyperglycemia in noncritically ill patients receiving parenteral and/or enteral nutrition. A comprehensive review of the literature assessing these patients noted superior glycemic outcomes with IV insulin titration compared with a subcutaneous approach. However, because IV insulin therapy poses several logistical challenges, subcutaneous insulin regimens using basal and supplemental rapid-acting insulin analogs have showed superior efficacy compared with an SS-only approach (16). The ADA recommends insulin coverage, including basal, prandial, and correctional components (7). Position statements of the Joint British Diabetes Societies elaborately outline the use of either insulin mixtures (e.g., premixed 70/30 insulin), isophane (NPH) insulin, the BB regimen already prescribed, or a BB regimen when on bolus feedings, as in Case 2 above (17). A recent article by Pasquel et al. (9) also recommends the use of basal insulin in conjunction with short-acting insulin every 4–6 hours.

Several challenges remain in managing glycemia in patients on enteral feedings, including the interruption of feedings for tests and procedures, with the ensuing risk of hypoglycemia. Literature also recommends starting a 10% dextrose infusion when enteral feedings are interrupted to prevent hypoglycemia (9,17).

Renal dysfunction increases the risk of hypoglycemia and adds a layer of complexity to diabetes management in the hospital setting. Fluctuating blood glucose levels surrounding variable inpatient dialysis schedules and dietary restrictions significantly influence treatment regimens. PD can have effects on glucose variability depending on the type of dialysis and dialysate used. Insulin remains the medication of choice for glycemic management in hospitalized patients on PD. It can be administered subcutaneously, or intraperitoneally along with the dialysate or directly into the intraperitoneal cavity. Some ambulatory studies suggest that insulin administered intraperitoneally may result in better glycemic outcomes (18). Most commonly, subcutaneously administered insulin is used in hospitalized patients because of its ease of administration and staff familiarity with it. Insulin pharmacokinetic profiles should guide product selection; however, recommendations are limited regarding which type of insulin is truly the most safe and effective. NPH insulin given at the start of nocturnal PD or regular insulin administered at the start of and 6 hours after PD initiation may also be used to stabilize glycemic trends while PD is running (19).

Herein, we have described a few distinct clinical scenarios to highlight the glycemic management challenges faced in the inpatient setting and the need for alternative strategies to the traditional BB regimens. This sentiment is becoming more widespread as inpatient diabetologists recognize the safety, efficacy, and ease of use of noninsulin agents (20). However, additional investigation will be imperative to validate the safety and efficacy of appropriate insulin and noninsulin treatments and to further develop inpatient guidelines that are applicable in real-world hospital settings.

No potential conflicts of interest relevant to this article were reported.

S.D., J.S., M.Z., and A.K. researched data, wrote the cases, and contributed to the discussion. B.P. and P.A. contributed to the introduction. A.R.S. reviewed/edited the manuscript. A.R.S. is the guarantor of this work and, as such, had full access to all the data included and takes responsibility for the integrity of the review.