A glucose range of 140 to 180 mg/dL is recommended for the majority of patients.

NICE-SUGAR¹ was a multinational randomized controlled trial across 42 different hospitals in which 6104 medical and surgical ICU patients were randomized to intensive (goal 81-108 mg/dL) versus conventional (goal <180 mg/dL) glucose control approaches. Eligible patients were adults who were expected to require care in the ICU for ≥3 consecutive days. Exclusion criteria included admission to the ICU for diabetic ketoacidosis (DKA) or hyperosmolar state, or prior hypoglycemia without full neurological recovery. The primary outcome was
death from any cause within 90 days. Secondary outcomes included hypoglycemia (≤40 mg/dL) as well as ICU and hospital length of stay.

The intensive control group had higher mortality (27.5% vs. 24.9% in the conventional control group, HR 1.11, 95% CI 1.01-1.23; P = .03) and more frequent hypoglycemia (6.8% vs. 0.5% in the conventional control group; P < .001). There were no differences between the groups in terms of ICU or overall hospital length of stay.

Based on the above study, the American Diabetes Association (ADA) recommends initiation of insulin for persistently elevated glucose levels ≥180 mg/dL with a target range of 140 to 180 mg/dL for the majority of critically and noncritically ill patients (grade A recommendation).²

Sole use of SSI is discouraged, as it results in higher blood sugars than basal-bolus insulin.

The RABBIT 2 trial³ was a multicenter, randomized trial comparing basal-bolus insulin to SSI alone in type 2 diabetic patients. One hundred thirty insulin-naive type 2 diabetic patients without ketoacidosis were randomized to receive basal-bolus insulin with glargine plus glulisine insulin or SSI alone. Patients in the basal-bolus group were started on a weight-based regimen with scheduled glargine and glulisine insulin with an additional glulisine SSI. Patients in the SSI group received regular insulin four times daily if eating or every 6 hours if fasting for blood sugar >140 mg/dL. Patients in the ICU, on systemic corticosteroid
therapy, or with known hepatic or renal disease (creatinine ≥3.0 mg/dL) were excluded. The primary outcome was mean daily blood glucose level. Secondary outcomes included frequency of hypoglycemia (<60 mg/dL), length of stay, and mortality.

Compared to those receiving basal-bolus, patients treated with SSI alone had higher mean daily (193 ± 54 vs. 166 ± 32 mg/dL; P < .001), fasting (165 ± 41 vs. 147 ± 36 mg/dL; P < .01), and random (189 ± 42 vs. 164 ± 35 mg/dL; P < .001) blood glucose levels. Compared to those in the basal-bolus group, fewer patients in the SSI group had mean glucose less than the 140 mg/dL target (38% vs. 66%; P-value not reported). Hypoglycemia occurred in two patients in each group (P-value not reported). Caveats of this trial include small sample size, limited treatment regimen options, and open-label protocol. ADA guidelines⁴ recommend basal-bolus insulin for those with adequate nutrition (grade A recommendation) and advise against prolonged use of SSI alone (grade A recommendation).

Discharge medication algorithms based on hemoglobin A1C can lead to improved outpatient glycemic control.

Two prospective, multicenter open-label studies have assessed the safety and efficacy of hospital discharge algorithms based on admission hemoglobin A1C. In the most recent study,⁵ patients who were already enrolled in the SITA-Hospital Trial (a multicenter randomized trial studying the effects of the addition of sitagliptin to inpatient insulin regimens) were invited to participate in a 6-month postdischarge study. Exclusion criteria included history of DKA, gastrointestinal obstruction,
corticosteroid therapy, and glomerular filtration rate (GFR)<30 mL/min/1.73 m². A total of 253 medical and surgical patients agreed to participate and were grouped according to their hemoglobin A1C level as measured during the hospitalization. Patients with a hemoglobin A1C <7% (controlled group) were discharged on either sitagliptin and metformin (and, if they were on insulin prior to admission, 50% of their inpatient insulin dose) or resumed their previous preadmission diabetic regimen; those with a hemoglobin A1C 7% to 9% (moderately uncontrolled group) were discharged on sitagliptin and metformin plus insulin glargine at 50% of their hospital insulin dose; those with a hemoglobin A1C >9% (severely uncontrolled group) were discharged on sitagliptin and metformin plus insulin glargine at 80% of their hospital insulin dose. Patients all received education on home glucose monitoring and hypoglycemia. Patients self-monitored glucose and were followed up by telephone every 2 weeks and clinic visits 1, 3, and 6 months after discharge. Coprimary outcomes were change in hemoglobin A1C at 3 and 6 months post discharge. Secondary outcomes included hypoglycemia (glucose <70 mg/dL) and “clinically important hypoglycemia” (glucose <54 mg/dL).