Improved safety with intravenous insulin therapy for critically ill patients with renal failure

Design

Retrospective cohort study 

N= 46

Objective

To evaluate the safety and efficacy of a new intravenous (IV) regular human insulin infusion (RHI) algorithm for glycemic control in critically ill patients with renal failure

Study Groups

Discontinued RHI algorithm (n= 21)

New RHI algorithm (n= 25)

Inclusion Criteria

Aged >18 years, admitted to study trauma center, referred to the Nutrition Support Service, required RHI infusion therapy; acute kidney injury (AKI), defined as a doubling of baseline serum creatinine concentration, decreased estimated glomerular filtration rate by 50%, or an hourly urine output < 0.5 mL/kg for 12 h or those with stage V chronic kidney disease (CKD; e.g., history of outpatient hemodialysis)

Exclusion Criteria

Methods

Patients' hospital charts were retrospectively reviewed, and eligible patients who were treated for hyperglycemia with the new RHI infusion algorithm and given concurrent continuous enteral nutrition (EN) or parenteral nutrition (PN) were compared with a historical control cohort who received the discontinued RHI algorithm. Time to achieve blood glucose (BG) control was determined from the difference between the hour of initiation of the RHI infusion to the hour whereby two consecutive hourly BG measurements were within 70 to 149 mg/dL. Please refer to the table below for graduated continuous IV RHI infusion algorithms:

Duration

Discontinued RHI algorithm (historical control): between February 2008 and May 2009

New RHI algorithm: between December 2009 and May 2012

Data collection: for a maximum of 7 d while receiving RHI infusion

Outcome Measures

Efficacy: mean daily BG, number of hours per day achieving the target BG range

Safety: hypoglycemia (moderate and severe episodes defined as BG concentration of 40-60 mg/dL and < 40 mg/dL, respectively)

Baseline Characteristics

New RHI algorithm (n= 25)

Age, years

Male, n

White

Diagnosis, n

Motor vehicle accident

Fall

Gunshot wound

Other 

7

5

5

4

11

5

4

5

History of diabetes mellitus

BMI, kg/m²

Serum creatinine, mg/dL

Predicted creatinine clearance, mL/min^*

Received hemodialysis

Sepsis

PN/EN

Length of stay, days 

Intensive care unit (ICU)

Hospital

27

38

31

44

^*p= 0.02

Results

Endpoint

Discontinued RHI algorithm (n= 21)

New RHI algorithm (n= 25)

p-value

Initial BG before RHI infusion, mg/dL

Time to achieve target BG, h

BG during infusion, mg/dL

Hours BG

< 70 mg/dL, h/d

70–149 mg/dL, h/d

70–179 mg/dL, h/d

> 179 mg/dL, h/d

1.3 ± 1.4

15.7 ± 3.3

20.5 ± 2.1

2.4 ± 1.6

0.2 ± 0.4

11.9 ± 2.5

16.3 ± 2.6

7.2 ± 2.7

Hypoglycemia 

BG 40–59 mg/dL

BG < 40 mg/dL

76%

29%

32%

0%

0.001

0.006

Duration of RHI infusion, d (9.2 ± 4.9 vs 9.0 ± 7.0) and amount of RHI received, units/g (113 ± 45 vs 102 ± 45) were not significantly different between the two groups. 

Adverse Events

See results; death: 4 vs 9 (not significant)

Study Author Conclusions

The new RHI algorithm improved patient safety by decreasing the prevalence of moderate hypoglycemia and eliminating severe hypoglycemia. The duration of glycemic control within the target BG range was decreased, but acceptable within a higher target BG ceiling. 

InpharmD Researcher Critique

Reducing the risk of hypoglycemia associated with intravenous insulin: experience with a computerized insulin infusion program in 4 adult intensive care units

Design

Prospective, observational, single-center study

N= 210

Objective

To evaluate blood glucose (BG) control in intensive care unit (ICU) patients treated with a computerized insulin infusion program (CIIP) in patients with end-stage renal disease (ESRD)

Study Groups

ESRD (n= 23)

Non-ESRD (n= 187)

Inclusion Criteria

ICU patients with 2 BG values >160 mg/dL; nil per os (NPO) or receiving continuous enteral tube feeds, or continuous parenteral nutrition

Exclusion Criteria

Diabetic ketoacidosis or nonketotic hyperosmolar syndrome

Methods

Data was collected from consecutive patients treated with IV insulin using the CIIP. The CIIP was programmed for a target BG of 140-180 mg/dL for medical ICU patients or 120-160 mg/dL for surgical ICU patients. Once beside nurse entered BG results into the CIIP, the program would periodically calculate the insulin infusion rate by using the formula (BG in mg/dL – 60) times a multiplier. The initial default multiplier is 0.02, which was changed in increments of 0.01 depending on the current BG readings (i.e., increasing if the current BG is above the target range or decreasing if the current BG is below the target range). The focus of this table is specific to data on the 11% of patients with ESRD.

Duration

January 1 to August 30, 2013

Outcome Measures

BG levels

Baseline Characteristics

Patients (N= 210)

Age, years

Male

Race

African American

Caucasian

Hispanic

Other

40%

44%

12%

4%

Inpatient HbA1c, %

6.9 ± 0.9

Estimated glomerular filtration rate, mL/min/1.73 m²

73 ± 44

End-stage renal failure

11%

ICU length of stay, days

9.6

Hospital length of stay, days

15

Results

Endpoint

ESRD (n= 23)

Non-ESRD (n= 187)

p-value

BG, mg/dL

Incidence of BG 60-79 mg/dL

Incidence of BG <70 mg/dL

Adverse Events

Common Adverse Events: Not disclosed

Serious Adverse Events: Not disclosed

Percentage that Discontinued due to Adverse Events: N/A

Study Author Conclusions

Hypoglycemia among ESRD patients was more common compared to non-ESRD patients, suggesting a need for a higher BG target specific to ESRD patient

InpharmD Researcher Critique

Management of Diabetic Ketoacidosis in a Patient With Chronic Kidney Disease Under Maintenance Hemodialysis: A Case Report

Design

Case presentation

A 47-year-old female with a complex medical history, including type 2 diabetes mellitus (DM), hypertension, and chronic kidney disease (CKD) due to chronic glomerulonephritis, presented with bilateral lower limb swelling, nausea, and altered sensorium. She had been undergoing maintenance hemodialysis twice weekly for four years and was on insulin therapy for nine years. Upon admission, the patient exhibited signs consistent with diabetic ketoacidosis (DKA) secondary to a urinary tract infection (UTI caused by E. coli).

Laboratory findings indicated significantly elevated blood glucose, creatinine, and urea levels, along with metabolic acidosis evidenced by a low pH and bicarbonate level. Initial management included fluid resuscitation with normal saline and potassium chloride, insulin infusion, intravenous sodium bicarbonate, and antibiotics for the UTI. Due to a decrease in urine output and rising potassium levels, dialysis was required.

Despite an episode of hypoglycemia post-dialysis, the patient's condition improved with continuous insulin infusion of 0.1 U/kg/h via pump. The insulin infusion was titrated per her blood glucose and potassium levels, prompting a 4-hour hold on the infusion when K dropped <3 mmol/L. The next day, her potassium levels rose, prompting hemodialysis. The patient experienced an episode of hypoglycemia following dialysis, which was treated with 50 mL of 50% dextrose. After a meal, her continuous insulin infusion was maintained at 0.05 U/kg/h. 

She was transitioned to alternate-day dialysis to manage fluid overload and uremic symptoms. Following stabilization of her metabolic acidosis and overall condition, she was successfully transferred from the ICU to the nephrology ward and eventually discharged.

Study Author Conclusions

The management of diabetic ketoacidosis (DKA) primarily involves insulin therapy, which effectively reduces serum glucose levels, inhibits the synthesis of ketone bodies, and corrects hyperkalemia. Intravenous insulin infusion is preferred over subcutaneous administration due to its ease of titration. For patients with DKA and end-stage renal disease (ESRD), insulin dosage typically requires reduction because insulin clearance is diminished when renal function is compromised. The optimal insulin infusion rate recommended is between 0.05 to 0.07 units/kg/h to avoid rapid declines in blood glucose and osmolality shifts. Monitoring can also be based on specific biochemical targets: a 0.5 mmol/h reduction in serum ketones, a 50-75 mg/dL/h decrease in serum glucose, and a 3 mmol/h increase in serum bicarbonate levels.

Patients with DKA and ESRD are more susceptible to hypoglycemia and other glycemic adverse effects than the general population, necessitating careful monitoring of capillary blood glucose levels. Kuverji et al. recommend a slower insulin infusion rate to mitigate hypoglycemic risks. In a reported case, a patient initially received six days of insulin infusion at 0.1 U/kg/h, later adjusted to 0.05 U/kg/h post-dialysis until achieving a blood glucose level below 200 mEq/L with corrected acidosis. Upon stabilizing and tolerating meals orally, the patient transitioned to subcutaneous regular insulin. The patient experienced a hypoglycemic episode after hemodialysis, which was managed with dextrose. Such hypoglycemic episodes can occur post-hemodialysis depending on the glucose concentration in the dialysate.