Hypoglycemic Agents and Anesthetic Techniques to Ameliorate Stress Hyperglycemia in Critically Ill Patients: A Systematic Review

Abstract

The concurrent utilization of hypoglycemic agents and anesthetic techniques has been demonstrated to mitigate stress hyperglycemia in critically ill patients without diabetes, thereby contributing to enhanced patient outcomes. Stress hyperglycemia, characterized by elevated blood glucose levels as a result of acute illness or physiological stress, frequently manifests in patients experiencing critical illness. This condition has been linked to augmented morbidity, protracted hospitalization durations, and elevated mortality rates. This review aims to introduce and critically assess various hypoglycemic agents and anesthetic techniques employed to alleviate stress hyperglycemia, emphasizing the necessity for continued research to comprehensively ascertain the safety and efficacy of these approaches, which will facilitate their broader integration.

INTRODUCTION

Elevated blood glucose levels, commonly referred to as hyperglycemia, frequently manifest in hospitalized patients without a pre-existing diagnosis of diabetes. This phenomenon is recognized as a constituent of the stress response accompanying various medical procedures and acute illnesses. Termed stress-induced hyperglycemia (SIH), this physiological response reflects the intricate interplay of stressors on glucose homeostasis in the absence of underlying diabetes. The presence of hyperglycemia in nondiabetic individuals during hospitalization underscores the intricate relationship between physiological stress and glucose regulation, highlighting the significance of understanding and managing SIH in the context of acute medical conditions. An increase in sympathetic activity secondary to stressful situations such as critical illness, surgery, or trauma causes the release of catecholamines, cortisol, growth hormone, and other counter-regulatory hormones that contribute to increased insulin resistance and subsequent elevations of blood glucose concentrations (BGCs).[1,2] This acute rise in BGC was, until 2001, believed to be a beneficial adaptive response that enhances the patients’ survival by providing a ready-to-use source of glucose during a time of increased demand. However, stress hyperglycemia is now recognized to be associated with an increased risk of infectious complications,[3,4,5] mortality in intensive and non-intensive care patients,[4,6,7,8] and extended hospital stays[4,5,9] by exacerbating the oxidative stress, altering lipid and carbohydrate metabolism, and disruption of integral pathways between the immune, endocrine, and the nervous systems.[10]

Based on the American Diabetes Association and the European Association for the Study of Diabetes (ADA/EASD) position statement and patients’ medical history, different types of hyperglycemia patients may experience during hospitalization can be categorized as shown in Table 1.[11,12,13]

Table 1.

Different types of hyperglycemia in hospitalized patients

Hyperglycemia type	Medical history	FBG (mg/dL)*	HbA1c (%)
Diabetic hyperglycemia
Diagnosed
Stringently controlled	+	FBG ≥126	HbA1c <6
Adequately controlled	+	FBG ≥126	6≤ HbA1c <7
Insufficiently controlled	+	FBG ≥126	HbA1c ≥7
Undiagnosed	−	FBG ≥126	HbA1c ≥6.5
Critical illness-associated hyperglycemia	−	FBG ≥126	HbA1c <6.5

*Or random BG ≥200 mg/dL. + means a medical history for diabetes is positive. − means a medical history for diabetes is negative. BG=Blood glucose, FBG=Fasting BG, HbA1c=Glycated hemoglobin

Such categorization may be essential as diabetes status impacts the relationship between glycemic control and clinical outcomes. Retrospective data suggest that patients with diabetes benefit less from glucose control than patients without diabetes,[14,15,16] and even more interestingly, a retrospective observational study demonstrated a reduction, rather than an increase, in mortality in patients with insufficiently controlled diabetes.[15] Furthermore, a multicenter observational cohort study of septic patients shows that increasing glucose variability is associated with an increased mortality rate; however, there is no association between increasing glucose variability and increased risk of death in patients with insulin-treated diabetes.[17] Another retrospective cohort study conducted to evaluate the relationship between diabetes status and the patient’s clinical outcomes shows that only hypoglycemia is independently associated with mortality risk in diabetic patients. In contrast, for patients without diabetes, both hyperglycemia and hypoglycemia are independently associated with a higher mortality rate.[18] These findings reveal that diabetes status modulates the relationship of the three domains of glycemic control (hyperglycemia, hypoglycemia, and glucose variability) with mortality; therefore, “one size fits all” is not the correct approach to control blood glucose in critically ill patients.

The era of glycemic control in critically ill patients began in the late 1980s by investigating the effects of treating diabetic patients with insulin in the cardiovascular surgery division of a hospital in Portland. Their findings demonstrated reduced mortality and infection rates of insulin-treated patients.[19,20] The interest in this field exploded with the publication of the landmark Leuven study[21] in 2001 that randomized 1548 adult surgical intensive care unit (ICU) patients and compared the conventional glucose management (infusing insulin when glycemia rose above 215 mg/dl and tapering down and stopping the infusion as soon as glycemia fell below 180 mg/dl) with continues insulin infusion strategy to target glycemia to 80–110 mg/dl. The latter approach became known as tight glucose control (TGC). TGC reduced mortality by 3%–4% and prevented many complications. A nonrandomized trial[22] and a less conclusively second Leuven trial[23] confirmed these findings. Two subsequent multicenter RCTs of TGC (Glucontrol[24] and VISEP[25]) were prematurely stopped due to demonstrating no benefit and a higher risk of hypoglycemia. Finally, the largest randomized controlled trial (RCT) of TGC, NICE-SUGAR, revealed another aspect of the TGC approach; Finfer et al.[26] reported higher 90-day mortality and a 13-fold increased risk of hypoglycemia-associated mortality with TGC than with standard care. In 2009, Griesdale et al.[27] conducted a meta-analysis regarding TGC’s influence on mortality and severe hypoglycemia rates in the ICU compared with conventional insulin therapy. Their study showed that TGC significantly increased the risk of hypoglycemia among critically ill patients admitted to a medical ICU without conferring any overall mortality benefit. However, this meta-analysis also demonstrated that TGC might benefit the patients admitted conclusively to a surgical ICU. A prominent finding supported by numerous studies revealed that insulin-induced hypoglycemia was associated with a higher risk of mortality rate in critically ill patients. Consequently, enthusiasm for the TGC approach has declined due to its elevated chance of hypoglycemia, and guidelines recommendations have been changed from “tight” glycemic control to “moderate” control.

Blood glucose control has remained a concern in different wards of hospitals, especially in ICU settings. Due to safety concerns, particularly after the NICE-SUGAR study,[26] scientists have been encouraged to find new strategies in critically ill patients to reach the optimum serum glucose range more safely (less hypoglycemia and hypokalemia) and more effectively (less hyperglycemia).

Numerous studies, which will be expounded on, have undertaken a comprehensive evaluation of the impact of various interventions, including long-acting insulins such as glargine and Neutral protamine hagedorn (NPH), short-acting insulin aspart, and alternative therapies such as metformin or incretin-based treatments, on the multifaceted dimensions of glycemic management protocols in critically ill patients. In addition, specific investigations have explored preoperative preparation modifications to prevent or mitigate postoperative hyperglycemia. Despite yielding beneficial effects on specific isolated outcomes, none of these strategies have fulfilled the criteria for unequivocal endorsement as the gold standard in controlling serum glucose levels in critically ill patients. The absence of a consensus recommendation reflects the ongoing challenge of comprehensively addressing the complex and diverse factors influencing glycemic control in this population, emphasizing the imperative need to identify further and refine aspects requiring improvement in the pursuit of optimal patient care.

This systematic review aimed to present various implemented strategies and agents for glycemic management in critically ill patients (except nutritional issues), comprehensively analyze their differences, and evaluate their safety and efficacy. The aim was to identify interventions that require further investigation to extend the evidence required to include them in the standard of care for hyperglycemia management in critically ill patients.

METHODS

In this systematic review, we searched online databases PubMed, EMBASE, Cochrane Central, Science Direct, Google Scholar, and the trial registry clinicaltrial.gov for RCTs published between database inceptions to July 21, 2021. We also set up “automatic NCBI searches and new record alerts” to find new articles on our topic of interest.

The search strategy consisted of several variations of terms related to the critically ill with SIH (participants), glycemic control (outcomes), and RCTs (study design) with no restrictions on the language. Keywords of each part were combined using Boolean operators.

([“Hyperglycemia” OR “insulin resistance” OR “glucose intolerance” OR “hyperinsulinism” OR “glucose metabolism disorders” OR “glycemic control” OR “blood glucose control”] AND [“critical illness” OR “critical care” OR “critically ill” OR “intensive care” OR “stress disorders” OR “stress disorders, traumatic, acute” OR “brain hemorrhage, traumatic” OR “brain hemorrhage, traumatic” OR “cerebral hemorrhage, traumatic” OR “brain stem hemorrhage, traumatic” OR “shock, traumatic”]).

In the formulation of our search strategy to investigate various facets of modalities designed to manage SIH in critically ill patients, we deliberately omitted the inclusion of specific interventions. Subsequently, on concluding the search process, all identified interventions were incorporated into our analysis, except supplemental or nutritional interventions. The rationale behind this exclusion stemmed from the extensive literature on these interventions, suggesting that a more focused and nuanced review of supplemental and nutritional strategies may be warranted in a distinct systematic review.

Reference lists of the included studies were also manually examined to identify any additional relevant studies.

We aimed to include all the studies with the following characteristics: (1) RCTs that reported at least one measurement of glycemic control as a primary outcome; (2) the population of interest was composed of adults (age >18 years) with SIH in critical illness or perioperative period without diagnosis of diabetes. Studies that explored the advantages of nutritional interventions, such as supplements or micro/macronutrients and the intensive insulin therapy approach in critically ill patients were excluded from our analysis. Furthermore, studies that were exclusively conducted on individuals with a pre-existing diagnosis of diabetes were also excluded from our systematic review.

After importing the retrieved studies and removing the duplicate records using a bibliography management software (EndNote^®), two reviewers independently screened the titles and abstracts for relevance in the first stage of study selection. Then, they extracted and selected eligible full-text and abstract-form studies. In the second stage, the full texts of the selected studies were independently examined by two investigators using a standardized eligibility form based on our pre-specified Population, Intervention, Comparison, Outcomes, and Study design (PICOS) criteria.

We extracted the following data from the included studies: First author’s last name, year of publication, study location, treatment and follow-up duration, patient population sample size, attrition rate, and our pre-specified PICOS information, including mean age, BMI, and gender of the study population, baseline glucose levels, description of the intervention employed and the reported outcomes using a standardized data sheet developed based on the Cochrane Consumers and Communication Review Group’s data extraction template. Table 2 (the appendix file) represents the characteristics of the studies and participants included in the systematic review.

Table 2.

Characteristics of included studies

Authors, year	Planned duration	Location	Sample size	Population	Description	Reported outcomes	Attrition rate (%)	Mean age (years)	Male (%)	BMI (kg.m−2)	Baseline glycemia	Follow up
Hypoglycemic agents
Insulin therapy protocols
Kim et al., 2012[34]	22 days	USA	32	Medical-surgical ICU	Intervention group: Glucose levels were monitored q4 h and subcutaneous rapid acting insulin (insulin aspart) administered based on both the previous insulin dose and current glucose level Control group: Glucose levels were monitored q1–2 h and patients were treated with the usual IVI protocol	Glucose level, insulin dose, workload of nurses, hypoglycemia rate	0/32 (0)	I: 62.4±3.2 C: 59.6±2.9	44	NA	NA	22 days
Bilotta et al., 2015[29]	-	Italy and Spain	36	Surgical and neurosurgical postoperative ICU	Intervention group: Insulin infusion was prepared by diluting 50 U Hlog in 500 mL of saline and infused through a volumetric pump Control group: Insulin infusion was prepared by diluting 50 U Hlin in 500 mL of saline and infused through a volumetric pump Crossover involved treatment of the same patient with both types of insulin (Hlog and Hlin) with a washout interval of at least 6 h between the 2 treatments when BGC values were >180 mg/dL	Mean BG values, lowest mean BG value, incidence of severe hypoglycemia (BG <90 mg/dL), mean infusion period, incidence of BG >180 mg/dL, extent of carry over effecta, duration of carry over, rate of BGC reduction, technical aspects	8/36 (22)	18–90 Most of the patients were between 51 and 90	36	26.56±3.55	194.75±16.25	-
Nader et al., 2020[30]	7 days	Iran	110	All ICU patients, candidates for insulin therapy	Insulin therapy was started when BG levels exceeded 180 mg/dL Intervention group: Continues infusion of insulin regular plus 15 IU insulin glargine subcutaneously Control group: Continues infusion of insulin regular	Daily variance in BG, mean daily BG, duration of time in target, insulin use, duration of BG >180 mg/dL, frequency of hypoglycemic events (BG <60 mg/dl) and severe hypoglycemia (BG <40 mg/dl), mechanical ventilation, ICU LOS, mortality	9/110 (7.6)	I: 60.2±14.9 C: 63.2±13.2	58	I: 27±3.7 C: 25.8±4.5	I: 224±47 C: 210±40	7 days
Metformin
Gore et al., 2003[37]	9 days	USA	10	Severely burned patients	Intervention group: The initial dosage of metformin was 850 mg. On the second postburn day, the 850-mg dosage was given twice at a 12-h interval. Beginning on postburn day 3, 850 mg of metformin was given three times per day at 8-h intervals. This 8-h dosage schedule was continued for the subsequent 6 days of study Control group; placebo - both groups: BG values of >220 mg/dL treated with insulin administered subcutaneously on a sliding scale basis	Endogenous glucose production during fasting, BG, glucose clearance, glucose oxidation, glucose uptake	0/10 (0)	I: 35±7 C: 38±3	NA	NA	I: 162±21.6 C: 203.4±30.6	28 days
Mojtahedzadeh et al., 2008[39]	7 days	Iran	33	SIRS positive multiple traumatized patients	In protocol A: Continuous IV infusion (50 IU insulin regular in 50 mL of 0.9% sodium chloride) In protocol B, patients received 1000 mg of metformin every 12 h In protocol C: Similar to protocol A, infusion of insulin started according to admission BGL. In addition, patients received 1000 mg metformin orally every 12 h	Glycemic control, insulin use, acid-base balance, lactate levels, mortality, APACHE score	4/33 (12.12)	A: 41.5±19.5 B: 47.5±14 C: 48.5±14.5	NA	NA	A: 191±28 B: 189±35 C: 192±28	7 days
Panahi et al., 2011[40]	3 days	Iran	52	SIRS positive trauma patients	Intervention group: 1000 mg q12 h Control group: Insulin 50 IU	BG, lactate and bicarbonate levels, APACHE score, GCS, blood pH	8/52 (15.4)	I: 48.4±16.77 C: 50±21.68	NA	NA	I: 177.81±20.17 C: 149.25±10.31	3 days
Mojtahedzadeh et al., 2012[41]	14 days	Iran	28	SIRS positive patients with serious injuries and major nonabdominal surgeries	Intervention group: 1000 mg q12 h Control group: Regular insulin	Glycemic control, general condition and organ function	4/28 (14.3)	I: 39.91±19.05 C: 43.16±11.67	70	NA	I: 177±22 C: 209±35	14 days
Metformin
Jeschke et al., 2016[38]	14 days	Canada	44	Severely burned adult patients	Intervention group: Metformin (500–1000 mg BID PO Control group: Insulin was started at 0.1 U/kg/h if the BG level exceeded 140 mg/dL and the infusion was adjusted to maintain euglycemia at 130 mg/dL according to a sliding scale	Glucose levels, insulin use, insulin and C peptide levels, inflammatory profile, mortality, insulin indices	0/44 (0)	47±18.3	73	NA	I: 122.4±7.21 C: 131.4±5.09	30 days
Incretin hormones
Dean et al., 2009[42]	270 min	Australia	7	Critically ill patients with different diagnosis	Patients were studied on 2 consecutive days Intervention group: GLP-1 for 270 min (i.e., t=0–270 min) Control group: Albumin for 270 min (i.e., t=0–270 min) At t=30 min a mixed nutrient liquid (64% carbohydrate, 1 kcal/mL), was delivered continuously into the small intestine (i.e., t=30–270 min)	BG level, insulin level, glycemic response to enteral nutrient, glucagon level. Insulin/glucose ratio	0/7 (0)	58	57	NA	NA	2 days
Lee et al., 2013[44]	420 min	Australia	20	Critically ill patients with different diagnosis	Patients were fasted for 6 h, and exogenous insulin was ceased 3 h prior to each study. They were randomly assigned to receive either the intervention (GIP) or control (IV 0.9% saline) in addition to IV GLP-1 at 1.2 pmol/kg/min Between T60 and T420, a mixed liquid nutrient was infused into the small intestine at 1.5 kcal/min	Glucose levels, insulin and glucagon levels	0/20 (0)	52±16	60	NA	126	2 days
Incretin hormones
Galiatsatos et al., 2014[45]	3 days	USA	18	Patients at burn and surgical ICU	After initiation of IIT protocol patients were randomly assigned to receive Intervention group: IV infusion of GLP-1 (7-36) amide for a continuous 72-h period Control group: IV infusion of saline for a continuous 72-h period	Glucose level, glycemic variability, insulin infusion rate, AUC0–72 h for the C peptide and glucagon, vasopressor use	0	I: 58.78±5.62 C: 67.11±4.37	78	I: 34.44±3.92 C: 34.12±2.96	NA	3 days
Kar et al., 2015[43]	360 min	Australia	24	Critically ill patients with different diagnosis	Intervention group: IV GIP Control group: Placebo (0.9% saline) 60 min after the study drug was commenced, a liquid nutrient meal containing carbohydrate (43%), fat (40%), and protein (17%), as well as 3 g 3-O-methyglucose dissolved in 5 mL water, 100 μg octanoic acid, and 20 MBq technetium-99 m calcium phytate was administered	Glucose levels, glucose absorption, insulin and glucagon concentration, gastric emptying	4/24 (17)	62	60	28	>128	2 days
Besch et al., 2017[46]	2 days	France	110	Post-CABG surgery patients	Intervention group: Exenatide (1 bolus of 0.05 μg/min followed by a constant infusion of 0.025 μg/min) Control group: Insulin (insulin lispro) 1 U/mL in a solution of saline	Number of patients spending at least 50% of the study period within the glycemic target, glucose levels, insulin use, glycemic variability	6/110 (5.45)	I: 68±11 C: 70±9	86	26.9	NA	2 days
Fayfman et al., 2018[47]	-	Two hospital in USA	96	Noncardiac surgery patients-without postoperative ICU admission requirement	Intervention group: 100 mg sitagliptin Control group: Placebo	Perioperative glycemic control, rate of hyperglycemia and hypoglycemia, ICU LOS, ICU transfer, complication	16/96 (16.7)	I: 51.1±12.8 C: 45.9±14.5	54	I: 28.3±6 C: 27.9±8.1	I: 107.6±255.3 C: 98.5±20	5 days
Incretin hormones
Cardona et al., 2019, RCT[48]	The day before surgery for up to 10 days	USA	60	Post-CABG surgery patients	Intervention group: 100 mg/day Sitagliptin Control group: Placebo	Perioperative glycemic control, rate of hyperglycemia and hypoglycemia, insulin use and duration of insulin therapy, ICU and hospital LOS, ICU transfer, perioperative complications	0/60 (0)	I: 64±11 C: 61±9	78	28±6	<126	Up to 10 days
Anesthesia techniques
Winterhalter et al., 2008[51]	During surgery	Germany	42	Cardiac surgery patients	Intervention group: Remifentanil was infused at 0.25 mg/kg/min following induction and for the entire duration of surgery. At sternotomy, a remifentanil bolus of 0.3 mg/kg was given Control group: Repeat bolus doses of 4 mcg/kg fentanyl every 30 min	Glucose level Stress hormones (epinephrine, norepinephrine, ADH, ACTH, cortisol, IL-6, IL-8, TNF-α), extubation time	0/42 (0)	I: 63±10 C: 64±7	83	I: 26.9±4.3 C: 27.9±3.4	I: 211±16 C: 130±43	20 h PO
IHN et al., 2009[49]	During surgery	Korea	40	Abdominal surgery-hysterectomy patients	TIVA group: Anesthesia was induced with remifentanil and propofol using a target controlled infuser VIMA group: The anesthetic circuit was primed with 6 vol% sevoflurane in 6 L/min N2O and 2 L/min O2 for 5 min	Glucose level Stress hormones (cortisol, adrenalin, noradrenalin) responses, intubation score, time to intubation and waking time, sore throat and postoperative nausea and vomiting	0/40 (0)	TIVA: 43.5±6.4 VIMA: 44.5±5.1	0	NA	>100	NA
Man et al., 2011[57]	30 min	Taiwan	60	Hysterectomy surgery patients	Intervention group: Receiving TENS on bilateral ST36 and SP6 acupoints for 30 min Control group: By contrast, the electrodes placed for the placebo group were not connected to the TENS stimulator	Glucose levels, insulin levels, HOMA index, mmean blood pressure, heart rate	8/60 (13.3)	I: 43.6±6.6 C: 44.4±8.2	0	NA	I: 96.84±15.84 C: 96.66±16.38	120 min
Anesthesia techniques
Taniguchi et al., 2013[53]	During surgery	Japan	46	Gastrectomy surgery patients	Group L: Remifentanil at infusion rate of 0.1 μg/kg/min Group H: Remifentanil at infusion rate of 0.5 μg/kg/min	HOMA-IR HOMA-β Glucose level Insulin level Stress hormones	9/46 (19.6)	L: 62.6±7.7 H: 59.7±6.7	70	NA	>100	12 h PO
Ibacache et al., 2020[54]	During surgery	Chile	40	Bariatric surgery patients	After anesthesia induction Intervention group: A bolus of 1 μg/kg dexmedetomidine in 10 min, followed by an infusion of 0.5 μg/kg/h until the end of surgery Control group: A bolus and infusion of 0.9% normal saline at the same rate as the intervention group until the end of surgery	Glucose levels, insulin levels	0/40 (0)	33±10.2	12.5	34.4±2.5	I: 80±15 C: 82±8	12 h PO
Subramaniam et al., 2020[52]	During surgery	USA	116	Cardiac surgery patients	Intervention group: Continuous remifentanil infusion at 0.1–0.4 mg/kg/min in the pre- and post-CPB periods. During CPB, the remifentanil IV infusion dose rate was increased to 1 mg/kg/min, but the dose rate could be decreased or the infusion stopped based on clinical judgment Control group: Intermittent fentanyl boluses of 50–250 mg during the surgery, based on clinician judgment and institutional standard of care	Intra and postoperatve glycemic control, insulin consumption, inflammatory profile	10/116 (8.6)	62.83±40	69	27.17±15.19	142±151.11	8 h PO
Others
Patel et al., 2014[58]	72 h	USA	104	Medical ICU patients	This study was a secondary analysis of a randomized controlled trial (n=104) of patients in the medical ICU randomized to receive physical and occupational therapy within 72 h of mechanical ventilation (early mobilization) or standard care with therapy as ordered by the primary care team	Median AUC glucose, daily ICU insulin, duration of delirium, ventilator-free days	0/104 (0)	I: 63±10 C: 64±7	83	I: 26.9±4.3 C: 27.9±3.4	I: 211±16 C: 130±43	28 days
Zhao et al., 2014[55]	-	China	40	Partial hepatectomy surgery patients	Intervention group (Group U): IV infusion of a total amount of 5000 IU/kg UTI before the induction of anesthesia and at the start of surgery Control group (Group C): Identical volume of physiological saline in the same manner	Fasting BG, insulin sensitivity index, insulin level, cortisol level, IL-6 level, glucagon level, hyperglycemia prevalence	0/40 (0)	51±10.5	77.5	23±2	NA	POD 2
Naderi-Behdani et al., 2022[56]	3 days	Iran	104	Surgical, medical, and traumatic critically ill patients	Intervention group: 6 mg melatonin BID for 3 days Control group: Placebo	Changes of blood sugar, Insulin resistance indices including homeostasis model assessment for insulin resistance and HOMA-AD ratios, GCS, ventilator dependency and delirium, mortality and ICU stay	8/104 (7.7)	56	64.6	I: 25.2±4.62 C: 26.12±5.2	>140	7 days

^aExpressed as the ratio between BGC reduction during insulin infusion (therapeutic BGC drop) and BGC reduction after infusion discontinuation (post-infusional BGC drop). ICU=Intensive care unit, SIRS=Systemic inflammatory response syndrome, CABG=Coronary artery bypass graft, IVI=Intravenous insulin infusion, SQ=Subcutaneous, BG=Blood glucose, BGC=BG concentration, BGL=BG level, PO=Postoperatively, GIP=Glucose-dependent insulinotropic peptide, GLP-1=Glucagon-like peptide 1, IIT=Investigator-initiated trial, TENS=Transcutaneous electrical nerve stimulation, IV=Intravenous, UTI=Urinary tract infection, APACHE=Acute physiology and chronic health evaluation, GCS=Glasgow Coma Scale, LOS=Length of stay, ADH=Antidiuretic hormone, ACTH=Adrenocorticotropic hormone, HOMA=Homeostasis model assessment, AUC=Area under the curve, HOMA-AD=HOMA adiponectin, TIVA=Total IV anesthesia, VIMA=Volatile induction and maintenance of anesthesia, NA=Not available, IL-6=Interleukin 6, IL-8=Interleukin 8, TNF-α=Tumor necrotizing factor-α, CPB=Cardiopulmonary bypass, POD 2=Postoperation day 2, BMI=Body mass index

We assessed the methodological quality of the included studies using the Cochrane quality assessment tool with the following domains: (a) random sequence generation, (b) allocation concealment, (c) blinding the participants and personnel, (d) blinding of outcome assessment, (e) incomplete data outcome, and (f) selective outcome reporting. Judgment about the risk of bias arising from each domain is generated based on the answers given to the signaling questions of a checklist (low risk, high risk, or unclear) [Table 3].

Table 3.

Risk of bias assessment

Article characteristics	Random sequence generation	Allocation concealment	Blinding of participants and personnel	Blinding of outcome assessment	Incomplete outcome data	Selective outcome reporting	Overall ROB judgment
Kim et al., 2012	High	High	High	Unclear	Unclear	Low	High risk
Bilotta et al., 2015	Unclear	Low	Low	Unclear	Low	Low	High risk
Nader et al., 2020	Low	Unclear	Low	Unclear	Low	Low	High risk
Gore et al., 2003	Low	Unclear	Low	Unclear	Low	Low	High risk
Mojtahedzadeh et al., 2008	Unclear	High	High	Unclear	Low	Low	High risk
Panahi et al., 2011	Unclear	Unclear	Unclear	Unclear	Low	Low	High risk
Mojtahedzadeh et al., 2012	Low	Unclear	Low	Unclear	Low	Low	High risk
Jeschke et al., 2016	Low	Low	High	Low	Low	Low	High risk
Dean et al., 2009	Unclear	Low	Low	Low	Low	Low	High risk
Lee et al., 2013	low	Unclear	Low	Unclear	Low	Low	High risk
Galiatsatos et al., 2014	Unclear	Low	Low	Unclear	Low	Low	High risk
Kar et al., 2015	Low	Low	Low	Unclear	Low	Low	Some concern
Besch et al., 2017	Low	High	Low	Low	Low	Low	High risk
Fayfman et al., 2018	Low	Low	Low	Unclear	High	Low	High risk
Cardona et al., 2019	Low	Unclear	Low	Unclear	Low	Low	High risk
Winterhalter et al., 2008	Low	Unclear	Low	Unclear	Low	Low	High risk
Ihn et al., 2009	Unclear	Unclear	Unclear	Unclear	Low	Low	High risk
Man et al., 2011	Low	Low	High	Unclear	Low	Low	High risk
Taniguchi et al., 2013	Low	Unclear	Unclear	Unclear	Unclear	Low	High risk
Ibacache et al., 2020	Low	Low	Low	Low	Low	Low	Low risk
Subramaniam et al., 2020	Low	Unclear	High	Unclear	Low	Low	High risk
Zhao et al., 2014	Low	Low	Low	Low	Unclear	Low	Some concern
Naderi-Behdani et al., 2022	Low	Low	Low	Low	Low	High	High risk

ROB: Risk of Bias

Due to the lack of similarity in population, dose or duration of interventions, target goals and outcomes across studies, and low methodological quality, a meta-analysis was deemed inconclusive.

Two independent investigators conducted the screening, data extraction, and risk of bias assessment. We tried to resolve any disagreements through discussion. A third investigator settled the discrepancies in the case of no consensus achievement.

RESULTS

The literature search provided 591 articles (including ten from manual searching). After the removal of the duplicates (260 records) and exclusion of the studies based on their abstracts or through their full-text examination (226 noneligible studies based on the inclusion/exclusion criteria; 38 non-RCTs; 32 pediatric studies; 4 animal studies; and 4 unpublished registered RCTs), a total of 27 were identified as eligible for inclusion [Figure 1]. Of the included studies, four papers were not published in English, and four were available only in abstract form. To present our findings better, we categorized them into three main groups: hypoglycemic agents, anesthetic techniques, and others.

Figure 1.

PRISMA flow chart

DISCUSSION

Most of the studies were carried out on a diverse population of critically ill patients or those undergoing surgery. A restricted number of studies examined the effects of specific interventions such as melatonin, dexmedetomidine, and ulinastatin, while variations in the duration of interventions were observed in metformin and glucagon-like peptide 1 (GLP-1) studies. Moreover, differences in comparison groups were seen in glucose-dependent insulinotropic peptide (GIP) studies, and the interventions employed in insulin therapy varied [Table 2]. These factors precluded the possibility of conducting a meta-analysis within each intervention group. Notably, the methodological quality of numerous studies was suboptimal, leading to their classification as high risk in terms of the overall risk of bias.

Insulin therapy

Insulin has several favorable metabolic and non-metabolic effects, reversing almost all undesirable adverse effects of stress response activation.[10] Therefore, insulin therapy is still (even after NICE-SUGAR) the principal approach to manage stress hyperglycemia. Inpatient hyperglycemia management frequently relies on the sliding scale of insulin (SSI) protocols. SSI protocols give a fixed insulin dose based on the current glucose level. Despite its simplicity, this approach may treat hyperglycemia after its occurrence rather than preventing it, leading to a rollercoaster effect (wide glycemic fluctuations) on glucose levels.

In contrast, with an algorithm, the insulin dose is adjusted over time to the actual requirements of the patient, eliminating the rollercoaster effect. The American Association of Clinical Endocrinologists and American Diabetes Association’s consensus statement on inpatient glycemic control suggests that continuous intravenous insulin infusion (IVI protocol) is the preferred route of insulin administration in critically ill patients.[28] After NICE-SUGAR, various approaches have been proposed to reduce IVI-associated hypoglycemia, including targeting higher blood glucose levels, more frequent blood glucose measurements, and increased caloric intake.

Bilotta et al.[29] conducted a multicenter, prospective, randomized trial to assess the impact of two distinct insulin infusion modalities in surgical and neurosurgical patients. The study utilized a crossover design, randomly assigning patients to undergo continuous Humulin insulin (Hlin) or Humalog insulin (Hlog) infusions facilitated by a volumetric pump. The primary endpoints of the investigation were characterized by evaluating the “carryover effect” associated with Hlog and Hlin. The “carryover effect” was quantified as the ratio between the reduction in BGC during insulin infusion (therapeutic BGC drop) and the decrease in BGC after infusion discontinuation (post-infusional BGC drop). In addition, the study examined the duration of the carryover effect, defined as the interval between insulin infusion discontinuation and the attainment of the lowest BGC value. The study’s findings revealed that treatment with Hlog, compared to Hlin, was linked to a less pronounced carryover effect and a shorter duration of carryover. This observation holds significance in addressing safety concerns related to continuous insulin therapy. Furthermore, Hlog demonstrated superiority over Hlin in reducing the number of BGC measurements and shortening the overall duration of insulin infusion.

It is noteworthy that the study focused exclusively on surgical and neurosurgical patients, and due to the crossover design implemented after the initial treatment (Hlin or Hlog), individuals who achieved target BGCs were subsequently excluded from further analysis. The outcomes of this investigation contribute valuable insights into the comparative efficacy and safety profiles of continuous insulin infusions using Humulin and Humalog in a specific patient population, thereby informing considerations for optimized glycemic management in surgical and neurosurgical contexts.

Insulin glargine, characterized by a stable plasma level and an extended half-life, holds promise for minimizing glucose level fluctuations in critically ill patients. Nader et al.[30] conducted a randomized trial to assess the impact of adding insulin glargine to routine continuous insulin infusion on glycemic control in critically ill patients. Their findings revealed that augmenting the protocol with insulin glargine led to lower glycemic levels. However, despite the absence of significant differences in blood glucose fluctuations between groups, the glargine cohort experienced more hypoglycemic episodes compared to the control group. Notably, glargine was utilized in the acute management of stress hyperglycemia in critically ill patients with unstable hemodynamic status, potentially contributing to the observed lack of significant differences in glycemic variability, which may be attributed to high doses of vasopressors for hemodynamic stabilization. Despite the increased incidence of hypoglycemia in the insulin glargine group, secondary outcomes such as ICU length of stay and duration of mechanical ventilation were lower with insulin glargine.

While insulin glargine is typically administered once daily, the variation in effective duration suggests that twice-daily dosing might enhance glucose control,[31] albeit with an elevated risk of hypoglycemia. Fox et al.[32] conducted a retrospective study supporting the safety and efficacy of twice-daily insulin glargine in stable critically ill patients.

Anderson et al.[33] demonstrated the comparable safety and effectiveness of insulin glargine and insulin NPH for hyperglycemia management in critically ill patients. Kim et al.[34] introduced a subcutaneous insulin algorithm (SQIA) as an alternative to intravenous insulin (IVI), achieving glucose goals while reducing the nursing burden with fewer glucose checks and less frequent insulin dosing. Huang’s study[35] showed that continuous subcutaneous insulin infusion resulted in better glucose control and a milder inflammatory response compared to multiple insulin injections in critically ill patients.

Noteworthy considerations include the diverse prevalence of diabetes in the patient population across studies, the variations in hypoglycemia definitions, and the utilization of point-of-care glucometers for measurements. Nader et al.’s comprehensive approach, involving twice-daily laboratory measurements, adds depth to understanding glycemic control in critically ill patients. Overall, these studies contribute valuable insights into the safety, efficacy, and practicality of insulin glargine and alternative glycemic management approaches in the complex context of critical care.

Metformin

The metabolic milieu in which hyperglycemia develops in critical illness in nondiabetic patients is complex and reflects stress response activation. As previously mentioned, a combination of different factors is involved in developing stress hyperglycemia. However, increased gluconeogenesis and hepatic insulin resistance are the key players in this metabolic disorder. Recent data suggest that increased hepatic glucose output may be more important than peripheral insulin resistance in stress hyperglycemia development.[36]

Metformin is the first-line medication for type 2 diabetes, which attenuates hyperglycemia primarily by reducing hepatic gluconeogenesis and increasing insulin sensitivity. The latter effect of metformin may allow the use of lower insulin doses in clinical settings and make an opportunity for critical patients to benefit from the nonmetabolic effects of insulin therapy while avoiding or minimizing the harmful effects of hypoglycemia due to the use of high insulin doses. Furthermore, studies in diabetic patients suggest that metformin is not associated with hypoglycemia, which is the principal concern associated with exogenous insulin therapy. Thus, it seems that metformin, whether as mono or adjunct therapy with insulin, may improve glycemic control in critical illnesses.

Gore et al.[37] conducted a placebo-controlled RCT to investigate the impact of metformin on BGC and glucose kinetics in severely burned nondiabetic patients. On the 7^th postburn day, with all patients receiving feeding, the metformin group exhibited significantly lower plasma glucose concentrations. None of the metformin-treated patients required exogenous insulin therapy, while two patients in the placebo group did. Metabolic assessments during fasting, intravenous glucose infusion, and a hyperinsulinemic-euglycemic clamp revealed that metformin administration led to reduced endogenous glucose production, accelerated glucose clearance, and increased insulin sensitivity. However, unlike prior studies in diabetic patients, this investigation reported increased insulin production.

Jeschke et al.[38] conducted a study administering metformin (500–2000 mg/day) for 14 days in severely burned nondiabetic patients, demonstrating a reduction in hypoglycemic episodes and lower insulin levels. While all control group patients required insulin therapy (for blood glucose >220 mg/dl), only one septic patient in the metformin group received insulin. However, daily mean glucose concentrations did not significantly differ between groups. Another study by Mojtahedzadeh et al. revealed a significant reduction in mean daily insulin requirements with metformin.[39]

In the study by Gore et al., patients had a higher mean burn size (71% vs. 37%) and a shorter time postburn injury. Three additional studies comparing metformin (2000 mg/day) with intensive insulin therapy (target blood glucose 80–110 mg/dl) in nondiabetic ICU-admitted patients favored metformin overall. However, BGCs did not significantly differ compared to insulin therapy, potentially due to the more intensive insulin therapy in these studies than those involving burn injury patients.

Panahi et al.[40] and Mojtahedzadeh et al.[41] demonstrated that metformin reduced mean blood glucose relative to baseline. The latter study reported a more profound decrease in the insulin-treated group. Interestingly, the incidence of hypoglycemia was zero in the metformin group in one study[40] and all groups in the two studies by Mojtahedzadeh et al.[39,41] Patients in Panahi et al. had higher APACHE scores than in other studies.

A notable concern regarding metformin use in the critically ill is its association with lactic acidosis. Although none of the mentioned studies reported significant differences in lactate levels or blood pH between comparison groups, cautious monitoring is essential due to the potential risks associated with renal dysfunction and tissue perfusion deficits in critically ill patients. Overall, while metformin shows promise in glycemic control, its use in critically ill populations requires careful consideration and vigilant monitoring due to potential complications.

Incretin hormones

GLP-1 and glucose-dependent insulinotropic peptide (GIP) are incretin hormones that attenuate high glucose levels in a glucose-dependent manner. This property makes incretin-based therapies a safe and promising intervention to treat hyperglycemia in hospital settings.

GLP-1 and GIP are rapidly metabolized by dipeptidyl peptidase IV (DPP-IV). Therefore, incretin-based therapies necessitate continuous infusion of exogenous GLP-1 or GIP, administration of a DPP-IV-resistant receptor agonist, or a DPP-IV inhibitor that increases endogenous GLP-1 and GIP concentrations.

Deane et al.[42] conducted a randomized, crossover study involving seven nondiabetic patients who received two intravenous infusions of GLP-1 and a placebo over 270 min on consecutive days. The study involved the infusion of a mixed nutrient liquid during the same period through a post-pyloric catheter. GLP-1 infusion reduced the overall glycemic response during enteral nutrition and peak BGC in all patients. In addition, GLP-1 administration increased the insulin/glucose ratio to 270 min. However, the study’s limitations, including a heterogeneous population, varied times of admission at the start of the study, and a small sample size, may impact the generalizability of the results.

In contrast, Kar’s study[43] demonstrated that the administration of GIP did not influence glycemia or insulinemia in critically ill patients. However, there was a significant rise in glucagon levels. The absence of a glucose-lowering effect of GIP in this study might be attributed to the glucagonotropic effect of exogenous GIP at normal to low blood glucose levels.

Lee et al.[44] investigated the additive insulinotropic effect of GIP when coadministered with GLP-1 in a randomized study. Their findings suggested that adding GIP to GLP-1 did not result in additional glucose-lowering or insulinotropic effects in critically ill patients. Glucagon levels also remained unchanged during the study. Despite using a lower dose of GIP compared to Kar et al.’s study, both studies included a heterogeneous population of nondiabetic critically ill patients admitted at different times relative to Kar’s survey.

Some studies explored the effects of GLP-1 receptor stimulation as an adjunct to intensive insulin therapy in critically ill patients. Galiatsatos et al.[45] assigned eighteen patients requiring insulin therapy to receive a 72-h continuous GLP-1 or normal saline infusion. Blood glucose control exhibited less variability in the GLP-1 group. Another study by Besch et al.[46] compared exenatide infusion with insulin therapy in nondiabetic coronary artery bypass graft surgery patients. Exenatide infusion reduced overall insulin consumption and increased the time to initiate insulin therapy, although most patients eventually received IV insulin for sustained hyperglycemia. However, the fasting status during the study period may have mitigated some of the beneficial effects of exenatide.

Two studies evaluated the effects of preoperative administration of sitagliptin on glycemic control.[47,48] The results indicated that sitagliptin was not potent enough to prevent stress hyperglycemia response in patients scheduled for surgery. The systemic immunological and metabolic responses induced by surgery and the fasting status before the procedure might have contributed to these outcomes. In addition, endogenous GLP-1 secretion stimulated by oral intake could have limited the impact of DPP-4 inhibition on GLP-1 levels.

In summary, the studies reviewed highlight the complex interactions between incretin hormones and glycemic control in critically ill patients. While GLP-1 administration shows promising effects in reducing glycemic response and improving insulin sensitivity, the role of GIP remains less clear, and coadministration with GLP-1 may not confer additional benefits. The use of incretin-based therapies as adjuncts to intensive insulin therapy presents varying outcomes, emphasizing the need for further research to elucidate their precise roles in different clinical contexts.

Anesthetic techniques

Surgery is a potent stimulant that induces endocrine, immunological, and metabolic responses, impairing glucose homeostasis. Various studies have been conducted to minimize the systemic response activation to improve patients’ outcomes. The anesthetic technique is one of the factors that may alter the intensity of the systemic response.

In their prospective, randomized trial, Ihn et al.[49] investigated anesthesia modalities in abdominal hysterectomy patients, employing a comparative analysis between two cohorts. Patients were randomly assigned to either the volatile induction and maintenance of anesthesia (VIMA) group, utilizing sevoflurane, or the total intravenous anesthesia (TIVA) group, receiving propofol and remifentanil. The research outcomes revealed noteworthy differences between the two groups’ stress hormone levels (adrenaline, noradrenaline, and cortisol) and glucose concentrations.

Specifically, the TIVA group exhibited significantly lower stress hormones and glucose levels than the VIMA group. This suggests that TIVA with propofol and remifentanil may contribute to a diminished stress response and glucose perturbations during abdominal hysterectomy procedures, implicating potential benefits in mitigating physiological stress.

Jung[50] conducted a study focusing on similar parameters in a related investigation within the same patient population. Notably, Jung’s findings indicated that the only statistically significant difference was observed in the stress hormone levels, with the VIMA group exhibiting higher levels at intubation than baseline. Furthermore, both the VIMA and TIVA groups demonstrated increased glucose levels during the extubation and recovery phases of the study, emphasizing the impact of anesthesia methods on metabolic parameters in this surgical context.

These findings collectively contribute to the growing body of evidence elucidating the physiological effects of different anesthesia techniques, providing valuable insights for clinical decision-making in abdominal hysterectomy procedures.

Previous studies also revealed that propofol and opioid combinations might prevent catecholamine release during surgery. High-dose opioid regimens are widely used in cardiac anesthesia due to their ability to preserve hemodynamic stability and attenuate the hormonal and metabolic response to surgical stress. However, a long postoperative ventilation period is usually required, and it increases hospitalization costs. Remifentanil is an ultrashort-acting opioid that would solve this problem due to its short half-life. Research has demonstrated that the continuous infusion of remifentanil, within the range of 0.1–0.4 μg/kg/min, exhibits efficacy in effectively suppressing inflammatory activation and stress hormone levels in patients undergoing cardiac surgery,[51,52] as compared to intermittent fentanyl administration. In a notable study by Subramaniam et al.,[51] a substantial reduction in stress hormone levels, fewer instances of hyperglycemia, and reduced insulin requirements were observed exclusively during the intraoperative phase in the group receiving remifentanil.

These findings suggest that the continuous administration of remifentanil within the specified dosage range may confer specific benefits in mitigating the inflammatory response and stress hormone release in the context of cardiac surgery. The observed reduction in hyperglycemia episodes and insulin requirements during the intraoperative phase underscores the potential impact of remifentanil in modulating physiological stress responses and metabolic homeostasis in this patient population. Such insights contribute to refining anesthesia strategies for cardiac surgery, offering valuable considerations for optimizing patient outcomes and perioperative care.

To ascertain whether the impact of remifentanil on postoperative insulin resistance is contingent on dosage, Taniguchi et al.[53] executed a randomized trial involving 37 patients undergoing elective gastrectomy. The participants were randomly allocated to two groups, receiving remifentanil at distinct infusion rates: 0.1 μg/kg/min (Group L) or 0.5 μg/kg/min (Group H). Analysis of their data revealed a significant reduction in postoperative insulin resistance in Group H as opposed to Group L, suggesting a dose-dependent relationship. Notably, the study reported a relatively elevated dropout rate.

In a related investigation, Ibachche et al.[54] explored the impact of dexmedetomidine in obese patients with impaired glucose tolerance undergoing bariatric surgery. Employing a bolus of 1 μg/kg administered over 10 min, followed by a continuous infusion of 0.5 μg/kg/h until the conclusion of the surgery, the researchers observed lower postoperative insulin levels compared to patients who received a saline infusion. Despite expectations of alterations in insulin secretion patterns with dexmedetomidine, no significant difference in postoperative blood glucose levels was observed between the two groups. These results imply that while dexmedetomidine may influence postoperative insulin levels, its effect may not be robust enough to produce discernible changes in BGCs in this patient population.

These studies contribute valuable insights into the nuanced relationships between drug dosage, perioperative metabolic responses, and the intricacies of glucose homeostasis. Further exploration and consideration of these findings may inform anesthesiologists and clinicians in refining strategies for managing insulin resistance and glucose regulation in surgical patients.

Others

In addition to the mentioned strategies, some other interventions were also effective for stress hyperglycemia management but could not be categorized in any of the mentioned categories.

Zhao et al.[55] conducted a study demonstrating that the administration of 5000 IU/kg ulinastatin, a broad-spectrum protease inhibitor, before the induction of anesthesia and at the start of surgery led to improved perioperative hyperglycemia and insulin resistance in patients undergoing partial hepatectomy. This improvement was attributed to the inhibitory effects of ulinastatin on inflammatory reactions associated with surgery.

In a study by Naderi-Behdani et al.,[56] melatonin supplementation (6 mg twice daily for three days) enhanced glycemia and reduced insulin resistance in a heterogeneous population of critically ill nondiabetic patients experiencing stress hyperglycemia. The beneficial effects were attributed to the anti-inflammatory properties of melatonin.

Man et al.[57] investigated the impact of perioperative transcutaneous electrical nerve stimulation (TENS) at bilateral ST36 and SP6 Chinese acupoints in patients undergoing hysterectomy. They found significant improvements in plasma glucose, plasma insulin levels, and the homeostatic model assessment for insulin resistance (HOMA) index at various time points after TENS discontinuation. It is important to note that the study included only female patients, raising concerns about the generalizability of the results to the broader population.

Exercise improves hyperglycemia in insulin resistance states and has anti-inflammatory effects. Patel et al.[58] conducted a secondary analysis of a RCT focused on early occupational and physical therapy in medical ICU patients with mechanical ventilation. They found that early mobilization significantly reduced insulin requirements to achieve glycemic goals (80–120 mg/dl). This suggests that incorporating exercise into the care of critically ill patients may contribute to better glycemic control.

In conclusion, the studies highlighted various interventions that have shown promise in improving glycemic control and reducing insulin resistance in critically ill patients. These interventions include using ulinastatin and melatonin to address perioperative hyperglycemia, perioperative transcutaneous electrical nerve stimulation, and early mobilization through exercise. However, the specific contexts, patient populations, and underlying mechanisms of action should be considered when interpreting and applying these findings.

CONCLUSION

Glycemic management in the context of stressful situations and critical illnesses presents a formidable challenge due to the myriad variables inherent in hospitalized patients. Nonetheless, it is unequivocal that hypoglycemia is correlated with an elevated risk of mortality, and intensive insulin therapy amplifies the likelihood of hypoglycemic events. This systematic review examines various interventions to enhance the effectiveness, safety, and technical aspects of conventional glucose management protocols in critical situations.

The utilization of SQIAs or long-acting insulins emerges as a potential avenue for achieving optimal glycemic control in the ICU and reducing the nursing burden through fewer glucose checks and less frequent insulin dosing. However, applicability to all patients is constrained, as subcutaneous insulin absorption may be suboptimal in certain conditions such as anasarca. Rapid insulin infusion demonstrates promise in attenuating blood glucose levels and diminishing the risk of hypoglycemia, particularly in patients exhibiting persistent hyperglycemia following initial insulin therapy. Notwithstanding the encouraging results, caution is warranted due to potential overestimation from the modest sample sizes. Validation of positive outcomes, such as the effects of insulin glargine on clinical endpoints in acutely unstable critical patients, necessitates future investigations encompassing larger cohorts with adequate statistical power.

Exploration of alternatives to insulin, including metformin and incretin hormones, has been undertaken in select small-scale studies. The efficacy of metformin has been assessed in a systemic inflammatory response syndrome-positive, heterogeneous ICU population, with varying durations of intervention and disparate blood glucose goals in studies examining more similar populations, such as trauma or burn injury. Concerns associated with the use of these agents in critically ill patients include an elevated risk of aspiration pneumonia linked to gastrointestinal side effects of GLP-1 and a heightened risk of lactic acidosis attributed to metformin usage. However, adequately powered studies examining the clinical outcomes of these insulin alternatives across diverse critical situations are currently lacking.

The consideration of incretin hormones, such as GIP infusion alone or in conjunction with GLP-1, appears less rational in critical illness, especially in fasting subjects, until clinical data delineate the effects of the duration and extent of hyperglycemia on GIP responsiveness in comparison to critically ill patients without antecedent hyperglycemia. Future studies evaluating incretin efficacy in SIH are encouraged to focus on more homogeneous and larger populations of critically ill patients.

For interventions discussed in the realms of “Anesthetic Techniques” and “Others,” the confirmation of positive effects necessitates further adequately powered randomized trials. These endeavors are indispensable for advancing our understanding of optimal glycemic management strategies in critical situations and fostering improvements in patient outcomes.

AUTHORS’ CONTRIBUTION

All authors contributed in the process of the study including conception and design, or acquisition of data, or analysis and interpretation of data; drafting and revising the manuscript for important intellectual content.

Financial support and sponsorship

This study was financially supported by Isfahan University of Medical Sciences.

Conflicts of interest

There are no conflicts of interest.