RABBIT 2

Published

Umpierrez GE, et al. "Randomized study of basal-bolus insulin therapy in the inpatient management of patients with type 2 diabetes (RABBIT 2 trial)". Diabetes Care. 2007. 30(9):2181-2186.
PubMed • Full text • PDF

Clinical Question

In non-critically ill inpatients with type 2 diabetes mellitus who were insulin-naive, is a basal-bolus insulin regimen more effective and safer for glycemic control as compared to sliding-scale regular insulin?

Bottom Line

In non-critically ill inpatients with type 2 diabetes mellitus who were insulin-naive, a basal-bolus insulin regimen is more effective for glycemic control as compared to sliding-scale regular insulin. There were no significant differences in the incidence of hypoglycemia and length of stay between the two treatment strategies.

Major Points

Poor glycemic control is associated with worse outcomes in critical and non-critically ill patients with diabetes mellitus (DM). At the time of the trial, the optimal management strategy for glycemic control of hospitalized patients was unclear. Although sliding-scale insulin (SSI) was widely used, there are significant disadvantages associated with this reactive strategy which attempts to treat rather than prevent hyperglycemia. Queale et al. reported that SSI is associated with a 3-fold increase in the risk of hyperglycemia as compared to patients not receiving glycemic therapy. In children admitted for diabetic ketoacidosis, SSI was associated with a longer duration of ketoacidosis and length of stay.

The Randomized Study of Basal-Bolus Insulin Therapy in the Inpatient Management of Patients With Type 2 Diabetes (RABBIT 2) trial compared the efficacy and safety of basal-bolus insulin regimen versus SSI for glycemic control of non-critically ill inpatients with T2DM who were insulin-naive. Patients treated with basal-bolus insulin had a significantly lower overall inpatient blood glucose difference, although there were no differences in the incidence of hypoglycemia or length of stay between the two strategies.

In 2011, the authors published a similar study on general surgical patients (RABBIT 2 Surgery), which again showed that basal-bolus treatment was associated with significantly better glycemic control and fewer complications as compared to SSI. In 2013, the authors published a study comparing basal-bolus, basal plus correction insulin regimen, and SSI for glycemic control of hospitalized general medical and surgical patients. Both basal-bolus and basal plus correction regimens resulted in significantly better glycemic control than SSI (P=0.04). However, there was a significant increase in the incidence of low blood glucose level (<70 mg/dL or <3.9 mmol/L) in the basal-bolus and basal plus correction regimens as compared to SSI (16% and 13% as compared to 3%, respectively; P=0.02). There was no difference in the incidence of severe hypoglycemia (<40 mg/dL or <2.2 mmol/L). A meta-analysis reported that non-SSI regimens are associated with significantly lower mean blood glucose level and lower incidence of hyperglycemia as compared to SSI. However, there were no significant differences in the risk of severe hypoglycemia or length of hospital stay between SSI and non-SSI treatment.

As a result of these and other studies, guideline-issuing bodies such as the American Diabetes Association advises against an SSI-only approach for hospitalized patients with DM, encouraging a basal-bolus approach.

Guidelines

ADA Standards of Medical Care in Diabetes, 2016

• A basal plus bolus correction insulin regimen is preferred for non-critically ill patients with poor oral intake or for those who are fasting. An insulin regimen with basal, nutritional, and correction components is preferred for patients with adequate nutritional intake. (grade A)
• The use of sliding scale insulin alone in inpatients is strongly discouraged. (grade A)

Design

• Multicenter, prospective, randomized, controlled trial
• N=130 non-critically ill inpatients with DM
  • Basal-bolus insulin (n=65)
  • SSI (n=65)
• Setting: 2 centers in Atlanta and Florida
• Follow-up: Length of hospital stay
• Analysis: Not specified
• Primary outcome: Glycemic control as measured by the mean daily blood glucose concentration

Population

Inclusion Criteria

• Age 18-80 years who were insulin-naive
• Non-surgical patients admitted to the general ward
• Known history of diabetes >3 months
• Blood glucose level between 140-400 mg/dl (7.8-22.2 mmol/L) on admission
• Diabetes treatment with diet alone or any combination of oral antidiabetic agents

Exclusion Criteria

• Active diabetic ketoacidosis
• ICU patients
• Corticosteroid therapy
• Expected to undergo surgery during hospitalization
• Clinically significant liver disease
• Serum creatinine ≥3.0 mg/dL (≥265 umol/L)
• Pregnancy
• Mental health condition causing the participant not able to understand the scope of the study

Baseline Characteristics

From the basal-bolus insulin group

• Demographics: Age 56±13 years, 64.6% male, 66.2% African-Americans
• BMI: 32±8 kg/m²
• Length of hospital stay 5.2±6 days
• Lab values:
  • Blood glucose on admission 229±71 mg/dl (12.7±3.9 mmol/L), mean blood glucose in hospital 166±32 mg/dl (9.2±1.8 mmol/L), mean fasting blood glucose 147±36 mg/dl (8.2±2.0 mmol/L), mean random blood glucose 164±35 mg/dl (9.1±1.9 mmol/L); HbA_1c 8.9±2
  • Creatinine 1±0.5 mg/dl (88.4±44.2 umol/L), hemoglobin 13±2 g/dL, white blood cell 9.6±4 ×10⁶

Interventions

• Patients were randomized to receive basal-bolus insulin or SSI
• Oral antiglycemic agents were discontinued on admission
• Blood glucose level was measured using a glucose meter before each meal and at bedtime (or every 6 h if a patient was unable to eat)
• The target fasting and pre-meal blood glucose level was <140 mg/dl (<7.8 mmol/L)
• Basal-bolus regimen:
  • For patients with blood glucose level between 140-200 mg/dl (7.8-11.1 mmol/L), total daily insulin dose of 0.4 units/kg was given
  • For those with glucose level between 201-400 mg/dl (11.2-22.2 mmol/L), total daily insulin dose of 0.5 units/kg was given
  • Insulin glargine (Lantus) was given once daily (50% of total insulin dose), insulin glulisine (Apidra) was given in 3 equal doses before meals (50% of total insulin dose)
  • If a patient was unable to eat, the pre-meal insulin glulisine was withheld but the insulin glargine was administered
  • If the fasting and pre-meal blood glucose levels were >140 mg/dl (>7.8 mmol/L), the daily dose of insulin glargine was increased by 20%
  • If a patient develops hypoglycemia (<70 mg/dl or <3.9 mmol/L), the daily dose of insulin glargine was reduced by 20%
  • Supplemental insulin glulisine was administered in addition to the scheduled pre-meal insulin if blood glucose >140 mg/dl (>7.8mmol/L)
• Sliding scale (SSI) regimen:
  • For patients who were unable to eat, regular insulin was administered every 6 hours, as per the “insulin sensitive” column
  • For patients able to eat, regular insulin was given before each meal and at bedtime, following the “usual” column
  • If fasting and pre-meal blood glucose levels are persistently >140 mg/dl (>7.8 mmol/L) without hypoglycemia, insulin scale was increased from the “insulin sensitive” to the “usual” column or from the “usual” to the “insulin-resistant” column
  • If a patient develops hypoglycemia (blood glucose <70 mg/dl or <3.9 mmol/L), insulin scale was decreased from “insulin-resistant” to “usual” column or from the “usual” to “insulin-sensitive” column
  • If the mean daily blood glucose level was >240 mg/dl (13.3 mmol/L), or if 3 consecutive levels were >240 mg/dl (>13.3 mmol/L) on the maximal sliding-scale regime, patients were switched to basal-bolus regimen starting with total daily dose of 0.5 units/kg

Outcomes

Comparisons are basal-bolus insulin vs. sliding-scale insulin

Primary Outcomes

Mean fasting glucose

147 vs. 165 mg/dL (P<0.01)

Mean random glucose

164 vs. 189 (P<0.001)

Mean glucose during hospital stay

166 vs. 193 (P<0.001)

Proportion of patients achieving target glucose level (<140 mg/dL)

66% vs. 38% (P not reported)

Secondary Outcomes

Hypoglycemia (≤60 mg/dL)

2 patients in each group

Hyperglycemia (≥240 mg/dL)

0 vs. 9 patients (P<0.05)

These patients were switched to basal-bolus insulin regimen and their glycemic control improved rapidly.

Mortality

1 reported death in the basal-bolus group due to pulmonary embolism

Total daily insulin dose

42 vs. 12.5 units (P<0.001)

Length of hospital stay

5.3 vs. 5.1 days (P=NS)

Criticisms

• Generalizability: Patients without documented DM before admission were excluded, and thus the trial likely disproportionately excluded patients with poor health literacy, poor access to care, or recent development of DM.
• Generalizability: Patients treated with insulin and corticosteroids were excluded due to the higher risk of hyperglycemia if they were treated with SSI.
• Clinical relevance: The study was not adequately powered to determine differences in mortality or clinical outcome between basal-bolus insulin regimen and SSI.
• Analysis unclear: There was crossover of 9 (14%) SSI patients to basal-bolus regimen; it is unclear if these were analyzed as intention-to-treat.
• Small sample size, conducted at only 2 sites, limits generalizability.
• There were significantly more males in the basal-bolus insulin group as compared to the SSI group (64.6% vs. 32.3%). However, the significance of this is unclear.
• The SSI group received a significantly less total daily insulin dose as compared to the basal-bolus group (12.5 vs. 42 units; P<0.001). It is unclear if the difference in insulin dose is related to the difference in outcomes observed between these treatments.

Funding

• Sanofi-Aventis. G.E.U
• American Heart Association
• National Institutes of Health
• General Clinical Research Center