What is the half life and/or duration of action of regular insulin when administered intravenously? At what time point may subcutaneous insulin be initiated following discontinuation of an insulin infusion?

Comment by InpharmD Researcher

Intravenous (IV) regular insulin has a short half-life, although its pharmacologic effects may persist for several hours. Reported half-life and duration of action vary according to the insulin product, dose, and administration conditions, with available prescribing information reporting half-life estimates ranging from approximately 20 minutes to 1 hour and pharmacologic effects lasting several hours in some evaluations. Current guidance favors initiating subcutaneous (SC) insulin before rather than after discontinuation of IV insulin. The American Diabetes Association (ADA) recommends approximately 2 hours of overlap for basal insulin and 1–2 hours of overlap when transitioning patients with resolved diabetic ketoacidosis (DKA) or hyperglycemic hyperosmolar state (HHS) to a multidose SC regimen, whereas the Joint British Diabetes Societies for Inpatient Care (JBDS) DKA guideline recommends continuing IV insulin for 30 minutes after administration of SC short-acting insulin. Furthermore, randomized trials evaluating early basal insulin administration during ongoing IV insulin therapy found shorter time to DKA resolution and reduced IV insulin requirements without significant increases in hypoglycemia or hypokalemia; however, these studies did not directly compare different timing or overlap regimens to determine an optimal approach.

Background

According to the 2026 American Diabetes Association (ADA) Standards of Care in Diabetes, when transitioning from intravenous to subcutaneous insulin, subcutaneous basal insulin should generally be administered 2 hours before the intravenous insulin infusion is discontinued to minimize rebound hyperglycemia while the subcutaneous insulin takes effect. The guideline also notes that administration of a low-dose basal insulin analog (0.15–0.3 units/kg) while intravenous insulin is ongoing may reduce insulin infusion duration, hospital length of stay, and rebound hyperglycemia without increasing the risk of hypoglycemia in patients with or without diabetic ketoacidosis. For patients with diabetic ketoacidosis (DKA) or hyperglycemic hyperosmolar state (HHS), the treatment algorithm recommends initiating a subcutaneous multidose insulin regimen after resolution and continuing the intravenous insulin infusion for 1–2 hours after subcutaneous insulin is administered. [1]

The 2022 Joint British Diabetes Societies for Inpatient Care guideline on the management of diabetic ketoacidosis in adults recommends continuing intravenous insulin and dextrose until the patient is eating and drinking normally. Long-acting basal insulin should be continued at the usual dose in patients already receiving it, while newly diagnosed patients should initiate long-acting basal insulin at 0.25 units/kg subcutaneously once daily. When transferring to subcutaneous insulin, the guideline states that the intravenous insulin infusion should not be discontinued until 30 minutes after subcutaneous short-acting insulin has been administered. [2]

A 2025 meta-analysis of 8 randomized controlled trials (N= 486) assessed early initiation of subcutaneous long- or ultra-long-acting basal insulin in combination with intravenous insulin infusion for diabetic ketoacidosis. Basal insulin was administered during intravenous insulin therapy, generally within 2–12 hours of diagnosis or treatment initiation, at doses ranging from 0.19–0.4 units/kg/day. Early subcutaneous basal insulin significantly reduced time to DKA resolution by 4.02 hours and total intravenous insulin use by 19.12 units, with no significant differences in rebound hyperglycemia, hypoglycemia, hypokalemia, length of stay, or in-hospital mortality. However, the included trials varied in basal insulin type, dose, and timing and did not directly compare different overlap intervals to determine an optimal timing strategy. [3]

Relevant Prescribing Information

Clinical Pharmacology [4]
Pharmacodynamics
With intravenous use, the pharmacologic effect of HUMULIN R begins at approximately 10 to 15 minutes and terminates at a median time of approximately 4 hours (range: 2 to 6 hours) after administration of doses in the range of 0.1 to 0.2 units/kg.

Pharmacokinetics
When administered intravenously, HUMULIN R had a mean half-life of approximately 20 minutes at a 0.1 unit/kg dose and 1 hour at a 0.2 unit/kg dose.

Clinical Pharmacology [5]
Pharmacokinetics
In a double-blind, randomized, crossover, euglycemic glucose clamp study, fifty eight healthy male subjects between 19 and 50 years of age received an intravenous infusion of either MYXREDLIN or another human insulin at 1 milliUnit/kg/min for 6 hours (0.36 units/kg total dose). On average, insulin concentrations of about 300 pM were attained approximately from 1.5 hours to 6 hours after starting intravenous infusion of MYXREDLIN, followed by a return to the baseline level 1.5 hours after stopping intravenous infusion.

Metabolism and Elimination
The mean terminal half-life from concentration data after stopping the intravenous infusion was estimated to be 23.4 minutes in healthy volunteers.

Clinical Pharmacology [6]
Pharmacodynamics
NOVOLIN R is a short-acting insulin. The time course of insulin action (i.e., glucose lowering) may vary considerably in different individuals, within the same individual, and different doses. When injected subcutaneously, the glucose-lowering effect of NOVOLIN R begins approximately 30 minutes post-dose, is maximal between 1.5 and 3.5 hours post-dose and terminates approximately 8 hours post-dose. The onset of action of NOVOLIN R, when administered intravenously, is more rapid in comparison to the subcutaneous administration. When injected subcutaneously, NOVOLIN R has a slower onset of action and longer duration of action compared to the rapid-acting insulin analogs.

Literature Review

A search of the published medical literature revealed 3 studies investigating the researchable question:

Level of evidence

C - Multiple studies with limitations or conflicting results Read more→

Please see Tables 1-3 for your response.

Maintaining Glycemic Control When Transitioning From Infusion Insulin: A Protocol-Driven, Multidisciplinary Approach
Design	Prospective study N= 114
Objective	To evaluate the effectiveness of a protocol-driven approach in maintaining glycemic control when transitioning patients from infusion insulin to subcutaneous insulin
Study Groups	Protocol followed (n= 66) Protocol not followed (n= 48)
Inclusion Criteria	All patients on continuous insulin infusion admitted to the surgical intensive care unit (SICU) or the medical/cardiac intensive care unit (MICU/CCU) between May 2008 and September 2008
Exclusion Criteria	Patients on infusion for less than 24 hours, had a liver transplant, discharged within 48 hours of transition, made comfort care, or transitioned to an insulin pump
Methods	A multidisciplinary hospitalist-pharmacist team implemented a three-step protocol for transitioning intensive care unit (ICU) patients from intravenous to subcutaneous insulin, with adherence left to provider discretion. Basal insulin was recommended for patients receiving diabetes medications, with A1c ≥6%, or receiving prednisone-equivalent doses ≥60 mg/day, provided the infusion rate was ≥1 unit/hour; patients with stress hyperglycemia received correctional insulin only. Total daily dose was calculated from the average infusion rate over the preceding 6 hours, adjusted for nutritional status, and divided into 50% basal and 50% nutritional insulin. Insulin glargine was administered 2 hours before infusion discontinuation, while nutritional insulin was given according to oral intake or tube feeding. Glucose values were assessed for 48 hours after transition and for up to 12 days in patients with diabetes.
Duration	May 2008 to September 2008
Outcome Measures	Primary: Mean blood glucose (BG) the first and second day after stopping the insulin infusion, number of patients with hypoglycemia (41-70 mg/dL) and severe hypoglycemia (<40 mg/dL) during the 48-hour transition Secondary: Severe hyperglycemia (>300 mg/dL), length of stay (LOS), re-initiation of the insulin infusion, day-weighted glucose mean 12 days following transition for patients with diabetes, identification of a new diagnosis of diabetes.
Baseline Characteristics		Protocol Followed (n= 29)	Protocol NOT Followed (n= 33)	Protocol Followed (n= 30)	Protocol NOT Followed (n= 9)
	Average age, years	57.7 ± 12.1	57.8 ± 12.3	56.5 ± 18.1	62.4 ± 15.5
	Male	72%	63%	66%	77%
	Body mass index, kg/m²	30.7 ± 7.2	28.6 ± 6.8	27 ± 5.4	25.2 ± 3
	History of diabetes	64%	86%	0	0
	Mean Hgb A1c	6.6% ± 1.2%	7.3% ± 1.8%	5.6% ± 0.3%	5.4% ± 0.4%
	Full nutrition	79%	61%	70%	100%
	On hemodialysis	17%	27%	10%	0
	On >60 mg prednisone or equivalent per day	24%	30%	0	0
Results		Protocol Followed (n= 33)	Protocol NOT Followed (n= 39)	p-value
	Average BG 0 to 24 hours post transition, mg/dL	167.98 ± 50.24	211.02 ± 81.01	<0.001
	Average BG 25 to 48 hours post transition, mg/dL	176.1 ± 55.25	218.2 ± 88.54	<0.001
	Total insulin used from 0 to 24 hours, units	65 ± 32.2	26.7 ± 25.4	<0.001
	Total insulin used from 25 to 48 hours, units	60.5 ± 35.4	28.1 ± 24.4	<0.001
	# of patients with severe hypoglycemia (<40 mg/dL)	1 (3%)	1 (2.6%)	*
	# of patients with hypoglycemia (41–70 mg/dL)	3 (9%)	2 (5.1%)	*
	% of BG values in goal range (80–180 mg/dL)	60.2% (153/254)	38.2% (104/272)	0.004
	# of patients with severe hyperglycemia (>300 mg/dL)	5 (15.2%)	19 (48.7%)	0.002
	*Too small of cell size to conduct chi-square analysis.
Adverse Events	Severe hypoglycemia occurred once in each group. Severe hyperglycemia was more common in the non-protocol group (49%) compared to the protocol group (15%).
Study Author Conclusions	Protocol adherence improved glycemic control, reduced unnecessary use of insulin, and identified patients with previously undiagnosed diabetes, without any increase in hypoglycemia.
Critique	This study provides a detailed, reproducible institutional transition process that integrates A1c testing, recent intravenous insulin requirements, nutritional status, pharmacist involvement, and administration of basal insulin before infusion discontinuation. However, the nonrandomized comparison based on provider adherence introduces selection and confounding risks, repeated transitions were analyzed for some patients, protocol adherence was only 58%, and the small nonadherent cohort without diabetes limits conclusions for patients with stress hyperglycemia.

Table 1 References:
[4] Ramos P, Childers D, Maynard G, et al. Maintaining glycemic control when transitioning from infusion insulin: a protocol-driven, multidisciplinary approach. J Hosp Med. 2010;5(8):446-451. doi:10.1002/jhm.810.

Intravenous Glargine and Regular Insulin Have Similar Effects on Endogenous Glucose Output and Peripheral Activation/Deactivation Kinetic Profiles
Design	Single-center, randomized, double-blind, crossover euglycemic glucose clamp study N= 15
Objective	To compare the effects of intravenously administered long-acting insulin analog glargine and regular human insulin on activation and deactivation of endogenous glucose output (EGO) and peripheral glucose uptake
Study Groups	All participants (N= 15)
Inclusion Criteria	Healthy male volunteers aged 27 ± 4 years, body mass index (BMI) 24.2 ± 0.7 kg/m2, no family history of diabetes, no significant medical problems, not on any medication affecting glucose metabolism
Exclusion Criteria	Not specified
Methods	Participants received intravenous insulin glargine or regular human insulin on two different study days in randomized order. During each glucose clamp, a 10-minute priming insulin dose was followed by a continuous intravenous infusion at 40 mU/m²/min for 4 hours, with a variable 20% dextrose infusion used to maintain euglycemia at 90 mg/dL. After the insulin infusion was discontinued, euglycemia was maintained for an additional 3-hour deactivation period. Endogenous glucose output and peripheral glucose disposal kinetics were evaluated during the 4-hour activation and 3-hour deactivation periods, including the time required for 50% deactivation of insulin effects after infusion cessation.
Duration	Not specified
Outcome Measures	Activation and deactivation kinetics of EGO and peripheral glucose uptake; absolute disposal rate
Baseline Characteristics		All participants (N= 15)
	Age, years	27 ± 4
	BMI, kg/m2	24.2 ± 0.7
	Fasting plasma glucose, mg/dl	89.0 ± 2.2
	Fasting insulin, µU/ml	8.1 ± 2.4
Results	All participants (N= 12)*	Regular Insulin	Insulin Glargine	p-value
	A₅₀EGO, min	73 ± 23	57 ± 20	NS
	R_{d max}, mg/kg/min	10.10 ± 0.77	9.90 ± 0.85	NS
	A₅₀IGDR, min	32 ± 5	42 ± 10	NS
	D₅₀IGDR, min	63 ± 5	57 ± 4	NS
	Abbreviations: A₅₀EGO, time required for 50% suppression of endogenous glucose output after insulin infusion; A50IGDR, time from basal Rd to half-maximum insulin-stimulated Rd; D50IGDR, time required for deactivation from maximum insulin-stimulated glucose disposal rate to half-maximum glucose disposal rate after cessation of insulin infusion; NS, not significant; R_{d max}, maximum rate of glucose disposal. *A total of 15 subjects participated; the first 3 were included in a pilot study to refine the methodology, and the remaining 12 subjects were included in the study results.
Adverse Events	Not specified
Study Author Conclusions	Activation and deactivation kinetics of EGO and peripheral glucose uptake as well as absolute disposal rate are similar between regular human insulin and insulin glargine when administered intravenously. Thus, the various biological actions of these insulin preparations when given subcutaneously are completely due to their different absorption kinetics.
Critique	The study's randomized, double-blind, crossover design allowed direct comparison between insulin types; however, the small sample of healthy male volunteers limits generalizability. The study provides pharmacodynamic data on regular insulin after infusion discontinuation, with insulin-stimulated glucose disposal decreasing to 50% of maximum at approximately 63 minutes.

Table 2 References:
[5] Mudaliar S, Mohideen P, Deutsch R, Ciaraldi TP, Armstrong D, Kim B, Sha X, Henry RR. Intravenous glargine and regular insulin have similar effects on endogenous glucose output and peripheral activation/deactivation kinetic profiles. Diabetes Care. 2002 Sep;25(9):1597-602. doi:10.2337/diacare.25.9.159

Improving Intravenous and Subcutaneous Insulin Overlap During Treatment of Diabetic Ketoacidosis: A Quality Improvement Project
Design	Quality improvement project N= 352
Objective	To reduce the frequency of insufficient overlap of intravenous (IV) and subcutaneous (SC) insulin during the treatment of diabetic ketoacidosis (DKA) as a quality improvement project
Study Groups	Preintervention (n= 251) Postintervention (n= 101)
Inclusion Criteria	Patients aged 18 years or older treated for DKA with IV insulin at a large tertiary care referral center in Rochester, Minnesota
Exclusion Criteria	Patients with b-hydroxybutyrate levels less than 3.0 mmol/L on admission
Methods	Rates of insufficient IV and SC insulin overlap were assessed in adults with DKA treated with IV insulin. An electronic medical record–based best practice advisory (BPA) was implemented to notify providers when discontinuing IV insulin if long-acting SC insulin had not been administered within the preceding 2–6 hours. The BPA reminded providers to overlap IV and SC insulin for at least 2 hours and provided guidance on discontinuing dextrose-containing IV fluids.
Duration	Preintervention: July 1, 2021, to September 14, 2022 Postintervention: September 15, 2022, to March 15, 2023
Outcome Measures	Primary: Rate of insufficient IV to SC insulin overlap Secondary: Rate of posttransition hypoglycemia, posttransition hyperglycemia, rebound DKA, length of hospital stay, duration of IV insulin therapy
Baseline Characteristics		Preintervention (n= 251)	Postintervention (n= 101)
	Age, years	47.5 ± 17.8	53 ± 20.1
	Female	126 (50.2%)	48 (47.5%)
	Male	123 (49%)	51 (50.5%)
	Type of diabetes - Type 1	153 (61%)	59 (58.4%)
	Type of diabetes - Type 2	83 (33%)	38 (37.6%)
	Admission b-hydroxybutyrate, mmol/L	6 ± 3.0	6.6 ± 3.3
Results		Preintervention (n= 251)	Postintervention (n= 101)	p-value
	Insufficient IV and SC insulin overlap	88 (35.1%)	20 (19.8%)	0.005
	Posttransition hypoglycemia	27 (10.7%)	4 (4.0%)	0.04
	Posttransition hyperglycemia	115 (45.8%)	39 (38.6%)	0.21
	Rebound DKA	17 (6.8%)	7 (6.9%)	0.97
Adverse Events	See above.
Study Author Conclusions	Using quality improvement methodology, the rates of insufficient IV to SC insulin overlap during treatment of DKA were measured and reduced through an electronic medical record-based best practice advisory targeting hospital providers.
Critique	The study effectively reduced the rate of insufficient insulin overlap (<2 hours) using a targeted intervention. However, it was limited by potential referral bias as it was conducted in a single tertiary care center, which may not represent other settings. Additionally, the BPA could contribute to alarm fatigue, and the study did not assess other aspects of DKA management, such as dosage of long-acting SC insulin or resolution of DKA before transition.

Table 3 References:
[6] Welch AA, Toro-Tobon D, Rachmasari KN, Sandooja RB, Rahimi L, Mohan S, Hewlett JR, Clark J, Maheshwari A, Zhang C, Brito JP. Improving Intravenous and Subcutaneous Insulin Overlap During Treatment of Diabetic Ketoacidosis: A Quality Improvement Project. Mayo Clin Proc Innov Qual Outcomes. 2024 May 22;8(3):293-300. doi:10.1016/j.mayocpiqo.2024.03.008