Early observations and pilot data that first suggested a new direction
Sliding-scale insulin (SSI) — reactive correction doses of regular insulin based on point-of-care glucose readings without basal coverage — was the dominant inpatient glucose management strategy for decades despite no evidence of benefit. Observational studies consistently associated both hyperglycemia and hypoglycemia with increased mortality, infections, and length of stay in hospitalized patients. Van den Berghe's landmark 2001 study in surgical ICU patients showed that tight glucose control (80-110 mg/dL) with intensive insulin therapy reduced mortality by 34% compared to conventional management. This single-center study catalyzed a global movement toward intensive inpatient glucose control, profoundly influencing ICU practice worldwide.
Landmark RCTs and pivotal trials that established the evidence base
The RABBIT 2 trial (2007) addressed the non-ICU setting, comparing basal-bolus insulin (glargine + glulisine) to sliding-scale regular insulin in 130 general medicine patients with type 2 diabetes. Basal-bolus achieved significantly better glucose control (mean glucose 166 vs 193 mg/dL) with a 38% higher rate of target glucose achievement and no increase in hypoglycemia. This established basal-bolus as superior to sliding scale for non-ICU inpatients. Then the NICE-SUGAR trial (2009) dramatically reversed the ICU tight control paradigm: in 6104 ICU patients, intensive glucose control (81-108 mg/dL) increased 90-day mortality by 14% compared to conventional control (≤180 mg/dL), primarily due to severe hypoglycemia. NICE-SUGAR was a watershed moment, demonstrating that the Van den Berghe findings did not generalize.
Follow-up studies, subgroup analyses, and real-world validation
Post-NICE-SUGAR, the field settled on moderate glucose targets (140-180 mg/dL for most ICU patients, 140-180 mg/dL for general wards). Research expanded the RABBIT 2 findings with RABBIT 2 Surgery showing similar basal-bolus superiority in surgical patients. The SITA-HOSPITAL trial demonstrated that sitagliptin (DPP-4 inhibitor) combined with basal insulin achieved similar glucose control to basal-bolus with fewer injections and less hypoglycemia, offering a simpler regimen. Continuous glucose monitoring (CGM) technology began transitioning from outpatient to inpatient use, with studies showing improved time-in-range and reduced hypoglycemia. The COVID-19 pandemic accelerated CGM adoption by reducing the need for frequent bedside fingerstick monitoring.
Integration into clinical practice guidelines and recommendations
ADA Standards of Care recommend glucose targets of 140-180 mg/dL for most hospitalized patients, with lower targets (110-140 mg/dL) considered for select patients if achievable without significant hypoglycemia. Sliding-scale insulin alone is strongly discouraged. Basal-bolus insulin is recommended as the standard subcutaneous regimen for non-ICU patients with diabetes. The Endocrine Society guidelines align, adding specific recommendations for insulin dose adjustment based on renal function, nutritional status, and glucocorticoid use. CGM use in hospitals is addressed with growing but still limited guideline support, pending larger outcome trials.
ADA Standards of Care — Diabetes Care in the Hospital
Target glucose 140-180 mg/dL for most hospitalized patients; basal-bolus insulin preferred over SSI alone; insulin is preferred pharmacological agent in hospital; avoid tight control (<110) based on NICE-SUGAR results
Endocrine Society Clinical Practice Guideline on Hospital Glycemia
Basal-bolus insulin for non-ICU patients; intravenous insulin infusion for ICU with glucose >180; target 140-180 for most patients; CGM as adjunct to POC glucose testing
Now
Current standard of care and ongoing research directions
Inpatient glycemic management has evolved from reactive sliding scale to proactive basal-bolus regimens with moderate targets. The major frontier is continuous glucose monitoring in the hospital, which provides real-time data and trend arrows that could enable more precise insulin dosing and earlier hypoglycemia detection. Several trials are evaluating CGM-guided insulin algorithms in both ICU and ward settings. Artificial intelligence-driven closed-loop insulin delivery systems (artificial pancreas) are being tested in hospital settings, potentially automating intravenous and subcutaneous insulin dosing. Despite progress, sliding-scale insulin remains stubbornly prevalent (used in >50% of non-ICU patients at many institutions), representing one of the largest evidence-practice gaps in hospital medicine.
Sliding-scale insulin is reactive (treats hyperglycemia after it occurs) rather than proactive (prevents it). It provides no basal insulin coverage, leading to wide glucose swings. The RABBIT 2 trial showed that basal-bolus achieves significantly better glucose control (mean 166 vs 193 mg/dL) without increasing hypoglycemia. Sliding scale perpetuates a cycle of hyperglycemia and catch-up correction that is associated with worse outcomes.
What happened with tight glucose control in the ICU?+
The initial Van den Berghe (Leuven) study showed a mortality benefit with tight control (80-110 mg/dL) in surgical ICU patients, but the larger multicenter NICE-SUGAR trial found that the same target increased mortality by 14%, primarily through severe hypoglycemia. The field now targets 140-180 mg/dL for most ICU patients, recognizing that the harm from hypoglycemia outweighs the benefit of tight normoglycemia in the critically ill.
Is continuous glucose monitoring ready for hospital use?+
CGM is increasingly used in hospitals but is not yet standard of care. Studies show improved time-in-range and reduced hypoglycemia, with COVID-19 accelerating adoption by reducing nurse-patient contact for fingerstick monitoring. Current guidelines support CGM as an adjunct to, not replacement for, point-of-care glucose testing. Regulatory approval for inpatient CGM use and larger outcome trials are needed before widespread adoption.
How should insulin be dosed when starting basal-bolus in a hospitalized patient?+
For patients eating regular meals: total daily dose 0.3-0.5 units/kg (lower end for elderly, renal insufficiency, or insulin-naive). Give 50% as basal insulin (glargine or detemir once or twice daily) and 50% as rapid-acting prandial insulin divided across three meals. Include correction-dose rapid-acting insulin with a sensitivity factor. Adjust daily based on glucose trends. Hold prandial doses if patient is NPO but continue basal insulin (reduce by 20-25% if needed).
