Evidence Against Sliding Scale
Relying on sliding scale insulin sometimes causes hypoglycemic events, severe hyperglycemic events, patient relapse after treatment of ketoacidosis, and the in-house development of diabetic ketoacidosis. In the study by Kathleen and colleagues, using sliding scale doses too high for patients with renal insufficiency was identified as the cause of six episodes of hypoglycemia.28
In the quality improvement project by Roman and colleagues, the use of quality indicators of a) BG < 40 mg/dL, b) BG > 450 mg/dL on two occasions, or c) BG > 45 with acetone, revealed that five cases were caused by physician failure to respond appropriately to hyperglycemia, despite administration of large amounts of additional regular insulin as coverage for capillary blood glucose (one of which resulted in the in-house development of ketoacidosis).29
In a retrospective review of management following ketoacidosis Gearhart and colleagues showed a higher median glucose among the patients treated with sliding scale than those treated proactively with insulin (prospective or anticipatory insulin), or treated with a combination of proactive insulin and “sliding scale” (correction dose insulin).30 Queale and colleagues showed that the use of either a standing dose of insulin or an oral hypoglycemic agent was associated with a reduced risk of hyperglycemic episodes, whereas sliding scale insulin regimens (when used without a standing dose of intermediate-acting insulin) were associated with an increased rate of hyperglycemic episodes.31
Baldwin and colleagues at Rush University Medical Center in Chicago recently reported the superiority of glycemic control among 88 patients for whom sliding scale was not allowed, in comparison with 98 well-matched controls from a historical comparison period.32 In the study group, standing orders for insulin were not permitted. House staff reviewed the results of glucose monitoring performed four times daily, and they twice daily reordered anticipatory insulin if appropriate. Glargine and rapid-acting analog were continued only for those patients already using such therapy prior to admission. Premixed insulins were not used. Oral agents were not used in combination with insulin and sometimes were discontinued for medical reasons.
In the control group, the percentages of patients with specific orders were: 100% sliding scale as defined by the authors, 32% twice-daily dosing with NPH/regular insulin, 37% orals agents, and 15% combination oral agents with NPH/regular insulin. In the study group, the percentages were: 0% sliding scale, 68% twice-daily NPH/regular insulin, 30% orals, and 0% combination NPH/regular with oral agents. For the study patients versus the historical controls, the mean blood glucose was 150 ± 37 mg/dL mg/dL versus 200 ± 51 mg/dL (p<0.01). The frequency of hypoglycemia between the two groups did not differ.
Schoeffler and colleagues in the St. Joseph’s/Candler Health System in Savannah, Ga., recently conducted a randomized study of 20 patients that reported the use of a 70/30 dose titration algorithm is superior to sliding scale insulin in achieving glycemic control.33 Patients were identified for possible enrollment through discovery of orders for sliding scale insulin in the pharmacy. After exclusion of patients receiving tube feeds or TPN and patients having type-1 diabetes, those consenting to be randomized either received the sliding scale as written by their physician or titration of 70/30 insulin given twice daily under algorithm. Hypoglycemia did not occur in either group. Downward trend over time and lower mean blood glucose (151.3 versus 175.6 mg/dL, p = 0.042) were observed in the 70/30 insulin group.