Hyperglycemia and Blood Glucose Control

Published on 22/03/2015 by admin

Filed under Critical Care Medicine

Last modified 22/04/2025

Print this page

This article have been viewed 1860 times

164 Hyperglycemia and Blood Glucose Control

Dieter Mesotten, Greet Van Den Berghe

Altered Glucose Regulation in Stress

At the end of the 19th century, Claude Bernard described the link between acute trauma and the development of hyperglycemia irrespective of underlying diabetes. It was considered to be an adaptive stress response ensuring adequate glucose supply to the obligatory glucose-consuming neurons, phagocytes, and reparative cells.^1,² Stress-induced hyperglycemia is evoked by integrated hormonal, cytokine, and nervous “counter-regulatory” signals on glucose metabolic pathways. Essentially, the hyperglycemia is due to insulin resistance in the liver and skeletal muscle. Hepatic insulin resistance leads to increased hepatic gluconeogenesis and glucose output.³ Decreased glycogen synthesis and a shift from insulin-dependent to non–insulin-dependent glucose uptake characterize skeletal muscle insulin resistance.⁴

In the acute phase of critical illness, it is assumed that increased levels of glucagon, cortisol, and growth hormone jointly increase hepatic gluconeogenesis. In addition, the catecholamines epinephrine and norepinephrine, released in response to acute injury, promote hepatic glycogenolysis. The cytokines interleukin (IL)-1, IL-6, and tumor necrosis factor (TNF) may directly or indirectly enhance both of these hyperglycemic responses.⁵

The important exercise-stimulated glucose uptake in skeletal muscle totally disappears because of the immobilization of the critically ill patient. Insulin-dependent glucose uptake is hampered also through a combined inhibition of glucose transporter-4 (GLUT-4) and glycogen synthase activity.^6,⁷ Although some studies have shown decreased glucose oxidation⁸ through pyruvate produced by glycolysis, others have demonstrated an opposite effect during critical illness.⁹ The decrease in insulin-dependent glucose uptake in skeletal muscle is completely offset by a strong increase in total body glucose uptake, of which the mononuclear phagocyte system in liver, spleen, and ileum are the main receivers.¹⁰ However, in skeletal muscle, non–insulin-dependent glucose uptake is also increased by increased expression of GLUT-1.^11,¹² The overall increased peripheral glucose uptake¹³ in light of hyperglycemia underscores the pivotal role of increased hepatic glucose production during critical illness, which cannot be suppressed by exogenous glucose.¹⁴

The position of adipose tissue in the regulation of glucose metabolism during critical illness has been neglected. Nevertheless, in diabetes mellitus, adipose tissue strongly modulates insulin resistance, as it is regarded as an insulin-dependent glucose uptake organ. Recent studies have now revealed that during critical illness, adipose tissue undergoes major changes.¹⁵ Possibly stimulated by illness-induced macrophage infiltration, adipocytes become more numerous and smaller and have an increased expression of the non–insulin-dependent glucose transporters, GLUT-1 and GLUT-3. The levels of GLUT-4 remain unaltered. As such, adipose tissue seems reprogrammed during critical illness to facilitate glucose uptake independent of circulating insulin levels.

Hyperglycemia in Critically Ill Patients

In a normal individual, blood glucose levels are tightly regulated within the narrow range of 60 to 140 mg/dL (3.3-7.7 mmol/L), both in fed and fasted states. Diabetic hyperglycemia is defined by the World Health Organization (WHO) as a fasting blood glucose concentration of 126 mg/dL (7 mmol/L) or higher and fed blood glucose levels higher than 200 mg/dL (11.1 mmol/L). In their 2006 guidelines, the WHO functionally defined normoglycemia as the glucose level associated with low risk of developing diabetes or cardiovascular disease. Unlike the diagnostic criteria for diabetes mellitus, no clear guidelines have been set for defining hyperglycemia in a critically ill patient. This explains the wide variations in the reported prevalence of hyperglycemia in critically ill patients.

However, stress hyperglycemia is also associated with adverse outcome in several critically ill patient populations. More precisely, a large cohort study of over 66,000 critically ill patients revealed a J-curved relationship between on-admission blood glucose level and the risk of mortality, with the nadir between 100 and 150 mg/dL (5.6-8.3 mmol/L).¹⁶ In patients with an acute coronary syndrome, a similar association has been observed, with the lowest risk of mortality at blood glucose levels between 80 and 100 mg/dL (4.4-5.5 mmol/L).¹⁷^–²⁰ Importantly, in patients with established diabetes mellitus prior to critical illness or an acute coronary syndrome, the relationship between hyperglycemia and mortality is significantly blunted and somewhat shifted to the higher blood glucose¹⁷ (Figure 164-1).

Figure 164-1 J-shaped association between blood glucose levels and mortality risk in critically ill patients. In patients without diabetes mellitus, hyperglycemia shows an almost linear relationship with mortality risk. Hypoglycemia is associated with an even steeper increase in mortality risk. Normoglycemia during critical illness conveys the lowest risk of dying. In patients with established diabetes, the J-shaped curve is significantly flattened out.

Until recently, it was considered state of the art to tolerate blood glucose levels up to 220 mg/dL (12 mmol/L) in fed critically ill patients. It was even suggested that this moderate hyperglycemia in critically ill patients was beneficial for organs such as the brain and the blood cells which rely solely on glucose for their energy supply and do not require insulin for glucose uptake. Motivation for treatment of blood glucose levels higher than 12 mmol/L was primarily the occurrence of hyperglycemia-induced osmotic diuresis and fluid shifts. Also, from the diabetes literature it was known that uncontrolled and pronounced hyperglycemia predisposes to infectious complications.²¹ In patients with known diabetes mellitus, usually more attention was paid to blood glucose levels and consequently more strictly controlled. This approach contrasts—in hindsight—with the blunting of the J-shaped relation between glycemia and mortality risk. Observational studies have also revealed that hyperglycemia in patients with established diabetes mellitus gives an at least threefold higher risk of mortality compared to patients with known diabetes.²²

Maintenance of Normoglycemia in the Intensive Care Unit

The Leuven Studies

In 2001, a large prospective, randomized, controlled trial (RCT) was the first to challenge the classic dogma of beneficial stress hyperglycemia.²³ It examined the effect of tight glycemic control (TGC) with intensive insulin therapy on mortality and morbidity of critically ill patients. Over a 1-year period, 1548 mechanically ventilated patients admitted to the intensive care unit (ICU), predominantly after extensive or complicated surgery or trauma, were randomly allocated to either intensive insulin therapy with blood glucose levels kept tightly between 80 and 110 mg/dL (4.5-6.1 mmol/L) or the conventional approach, which recommended insulin therapy only if blood glucose levels exceeded 12 mmol/L. The intervention of TGC comprised accurate arterial blood glucose measurements by a blood gas analyzer and a reliable continuous infusion of insulin exclusively via a central venous line, using an accurate syringe-driven infusion pump. The fine insulin dose adaptations were performed by trained bedside nurses and based on a guideline which requires a high level of intuitive and anticipating decision making. In this study, patients were kept in a nonfasting state at all times. Dextrose 20% was administered on the first day (192 g glucose over 24 hours or 768 kcal/d). Thereafter, enteral nutrition was started, with the daily amount progressively increased as tolerated. When enteral nutrition was insufficient, early supplemental parenteral nutrition was given, resulting in administration on average of 1100 nonprotein kcal/d.

Intensive insulin therapy, resulting in the administration of on average 1100 nonprotein kcal/d, lowered ICU mortality from 8% to 4.6% (absolute risk reduction [ARR] 3.4%) and in-hospital mortality from 10.9% to 7.2% (ARR 3.7%). The effect occurred particularly in the population with prolonged critical illness, among whom mortality was reduced from 20.2% to 10.6%. Even patients in the conventional insulin treatment schedule with only moderate hyperglycemia (110-150 mg/dL) showed higher mortality compared with patients in the strict glycemic control schedule.²⁴ Intensive insulin therapy also had a major effect on morbidity. It decreased the duration of ventilatory support and ICU stay, reduced the need for blood transfusions, and lowered the incidence of bloodstream infections and excessive inflammation. Even more striking, intensive insulin therapy caused a highly significant decrease in the development of critical illness polyneuropathy and acute kidney failure.

Subsequently, the effect of TGC was tested in a medical ICU setting by the same group.²⁵ The difference in in-hospital mortality, 40.0% in the control group and 37.3% in the intervention group, was not statistically significant in an intention-to-treat analysis of the 1200 included patients. However, in patients who stayed in the ICU for 3 or more days, in-hospital mortality was reduced from 52.5% to 43.0% by TGC. Intensive insulin therapy also reduced morbidity (incidence of acute kidney failure, weaning of the ventilator, ICU/hospital stay) but not as strikingly as in the surgical study. This was in part explained by a larger fraction of patients in medical ICUs who were admitted with established organ damage, possibly reducing the opportunity of prevention by glucose lowering.²⁶ The fact that intensive insulin therapy to normal-for-age blood glucose targets in mainly postoperative pediatric critically ill patients did reduce mortality by an ARR of 3% may further corroborate this finding.²⁷

The downside of TGC has been the increase in the incidence of hypoglycemia (blood glucose levels < 40 mg/dL [<2.2 mmol/L]) despite improving patient outcome. In the Leuven studies, 5.1% (surgical ICU), 18.7% (medical ICU), and 25% (pediatric ICU) of patients randomized to TGC experienced at least one episode of hypoglycemia. To date, long-term follow-up studies to gauge the impact of brief hypoglycemia on neurocognitive function have been lacking. In addition, it is possible that fluctuations in glucose levels such as those induced by insulin therapy based on inaccurate glycemic monitoring, or by overcorrection of hypoglycemia, may be more deleterious that hypoglycemia by itself. Such aspects remain to be investigated in great detail.

The Initial Repeat Studies

Two European multicenter studies designed to assess whether intensive insulin therapy exerts benefit, with mortality as the primary endpoint, failed to reproduce the Leuven findings. The VISEP (Volume substitution and Insulin therapy in severe SEPsis) (N = 537) trial was designed as a four-arm study to assess the difference between two choices of fluid resuscitation (10% pentastarch versus modified Ringer’s lactate) and the efficacy and safety of intensive insulin therapy in patients with severe sepsis and septic shock.²⁸ In this study, blood glucose targets comparable to the Leuven studies were set out for the intervention (80-110 mg/L) and control (180-200 mg/dL) groups. Likewise, the insulin administration and blood glucose measurements had been standardized. Nevertheless, the insulin arm of the study was stopped early after 488 patients had been included, because the rate of hypoglycemia (12.1%) in the intensive insulin therapy group was considered unacceptably high and may be associated with higher mortality. Then at the first planned interim analysis, the fluid resuscitation arm of the study was also suspended because of increased risk of organ failure in the 10% pentastarch arm. The primary endpoint, 90-day mortality, was 39.7% in the intensive versus 35.4% in the conventional treatment arm.

The GLUCONTROL multicenter RCT (N=1101) investigated whether tight glycemic control (80 and 110 mg/dL) with intensive insulin therapy versus an intermediate target for blood glucose (140-180 mg/dL [7.8-10.0 mmol/L]) improves survival in a mixed population of critically ill patients.²⁹ This study was also stopped early because the target glycemic control was not reached and the incidence of hypoglycemia was 9.8%. ICU mortality did not differ between the intensive insulin therapy group (17.2%) and the control group (15.3%).

Two single-center studies in a mixed medical/surgical ICU population, both smaller than the Leuven studies, followed and were unable to reproduce a significant mortality benefit.^30,³¹ In contrast, a number of small RCTs in selected subpopulations, mostly focusing on morbidity as primary endpoint, as well as several larger implementation studies revealed improved outcome as did the Leuven studies.³²^–³⁵

Nice-Sugar

All the described studies were in fact statistically underpowered to detect a reasonable mortality difference. To address this issue, the NICE-SUGAR (Normoglycemia in Intensive Care Evaluation and Survival Using Glucose Algorithm Regulation) included 6100 patients over 41 participating centers.³⁶ This study compared a blood glucose target of below 108 mg/dL (<6.0 mmol/L) versus “usual care,” which meant an intermediate blood glucose target of 140 to 180 mg/dL (8 to 10 mmol/L). Owing to the results from the Leuven studies, clinicians had become aware of the negative impact of hyperglycemia, so tolerating higher glucose levels was considered unacceptable or even unethical by clinicians and investigators. The aim of NICE-SUGAR, therefore, was to assess whether further lowering blood glucose levels to less than 108 mg/dL (<6.0 mmol/L) in a broad context of clinical practice in ICUs, predominantly located in Australia and New-Zealand, and using the normal daily clinical practice tools available would exert additional benefit. Contrary to expectations, NICE-SUGAR revealed that targeting 108 mg/dL with insulin increased 90-day mortality from 24.9% to 27.5% as compared with the 140 to 180 mg/dL (8-10 mmol/L) glucose target. Excess deaths were attributed to cardiovascular causes.

Coiitss

Patients with septic shock requiring administration of glucocorticoids are faced with a high mortality risk; the severity of illness and glucocorticoid treatment make hyperglycemia common. Therefore, this would be an optimal population in whom to study whether TGC could reduce mortality. In the Corticosteroids and Intensive Insulin Therapy for Septic Shock (COIITSS) multicenter study, 509 patients were randomized to either intensive insulin therapy aiming for blood glucose levels between 80 and 110 mg/dL or to conventional insulin therapy.³⁷ In the latter group, an intermediate target was used, as the physicians were recommended to follow the 2004 Surviving Sepsis Campaign Guidelines (blood glucose levels < 150 mg/dL [8.3 mmol/L]). Hospital mortality in the intensive insulin therapy group (45.9%) did not differ from the conventional group (42.9%). Poor separation of the blood glucose levels between the study groups and the small size of the study may have made it hard to detect any treatment effect of TGC.

Meta-Analyses

Nowadays, practice guidelines ideally are based on systematic reviews and meta-analyses. The two most recent meta-analyses showed that in critically ill adult patients, TGC did not significantly reduce hospital mortality but is associated with an increased risk of hypoglycemia.^38,³⁹ However, TGC may be beneficial to patients admitted to a surgical ICU.

Critical Appraisal of the Evidence for Tight Glycemic Control in the ICU

Given that the effect of controlling blood glucose levels during critical illness ranges from benefit, to no effect, to potentially harmful, most clinicians are now in agreement that blood glucose levels do in fact play a role in patient outcome. The pre-2001 era where blood glucose levels were hardly measured in critically ill patients has passed forever. However, discrepancies in the study results have made it difficult to make strong recommendations. Likewise, consensus statements on glycemic management of hospitalized patients by the American Association of Clinical Endocrinologists and the American Diabetes Association have changed significantly over the last years.⁴⁰ While the 2004 and 2006 statements recommended stricter targets for glycemic management in the ICU, in 2009 it was advised that the starting threshold for intravenous insulin therapy in the ICU should be 180 mg/dL (10 mmol/L). And once started, blood glucose levels should be maintained between 140 and 180 mg/dL (7.8-10 mmol/L). Somewhat lower levels may be appropriate in selected patient populations. Targets below 110 mg/dL (<6.1 mmol/L) are not recommended.

Still, more can be learned from the differences between the Leuven proof-of-concept studies and the subsequent repeat trials.⁴¹ First, “normoglycemia” was compared with distinct “control” targets (140-180 mg/dL or 8-10 mmol/L in NICE-SUGAR and GLUCONTROL; 180-215 mg/dL or 10-12 mmol/L in Leuven), making the studies fundamentally different. The control group in the Leuven studies reflected the assumption of hyperglycemia as a potentially beneficial adaptation. Hence, a “do-not-touch” approach unless glucose exceeded the renal threshold of 215 mg/dL was used in this group. In contrast, the NICE-SUGAR trial was executed in the “flatter” part of the observational glycemia-mortality risk curve, with 70% of the patients in the control group receiving insulin treatment to target an intermediate blood glucose level of 140 to 180 mg/dL (8-10 mmol/L) (Figure 164-2).¹⁶^–¹⁹ The control group in NICE-SUGAR, as a result of the changed usual care, already could have benefited from reducing blood glucose as compared with the control group in Leuven. The lower observed mortality than the carefully documented expected mortality (24.9% versus 30%, respectively) in the NICE-SUGAR control group may indeed suggest that there was already such a benefit in the control group.

Figure 164-2 Comparing tight glycemic control with “usual care” strongly depends on mortality risk in the control group. In the Leuven proof-of-concept studies, tight glycemic control was compared with the usual care of tolerating hyperglycemia up to the renal threshold (215 mg/dL). The control group in the NICE-SUGAR trial targeted intermediate blood glucose levels (145-180 mg/dL).

Second, the level of therapy compliance, in this case the degree of success in reaching and maintaining the preset target range for glucose in the intervention group, as well as the degree of overlap with the control group, varied greatly between the studies. The methodological aspects of glucose measurement and the level of expertise of the nursing team with blood glucose control in the Leuven studies may have played a key role. In the Leuven studies, 70% of the patients in the intervention group were on average in target,⁴² whereas this was much less than 50% in NICE-SUGAR and in several of the other repeat studies. This could be important, as a recent meta-analysis suggested that studies that actually managed to adequately achieve the blood glucose target showed a reduced mortality, whereas studies that did not succeed in reaching the target reported no benefit or even increased mortality.^38,⁴³ Maintaining normoglycemia may be more feasible in patients after surgical critical illness than in those with medical illnesses.

Third, a requirement for safe insulin dose adjusting to reach and maintain normoglycemia is a standardized, accurate glucose measurement technology. In NICE-SUGAR, a variety of glucose meters were allowed, whereas most of them have recently been shown to be unsuitable for this purpose.⁴⁴ Accuracy of certain glucometers has been shown to be extremely poor in the ICU setting, and the wide error goes in the opposite direction for the low and high glucose ranges, making it impossible to use them for targeting a very narrow glucose range.^45,⁴⁶ In addition, varying sampling sites (arterial, venous, and capillary) were accepted in the context of routine clinical practice, and these too have led to erroneous results for blood glucose.⁴⁷ Inaccuracy of glucose measurement may have misguided the insulin titration and thereby induced (undetected) hypoglycemia and large blood glucose fluctuations. Avoiding highly variable blood glucose levels requires experience and thus has a learning curve, which is inherent with complex interventions.

Fourth, feeding strategies differed in the major studies. The substantially higher amounts of parenteral nutrition in the Leuven studies, although still on average below normal caloric requirements, may have increased the severity of stress-induced hyperglycemia, and thus the intervention may have been in part directed to counteract this side effect of parenteral nutrition. In NICE-SUGAR, feeding relied almost exclusively on the enteral route (80 kcal intravenous glucose on the first day; on average a total of 880 kcal/d), whereas in Leuven, early parenteral nutrition (768 kcal on the first day) supplemented insufficient enteral feeding, resulting in an average 1100 kcal/d for adult patients. Insulin treatment in a nutritionally deprived state early in the disease course, as in NICE-SUGAR as a result of their feeding guidelines, may have been deleterious by evoking a global substrate deficit via insulin-induced counteracting of proteolysis, lipolysis, glycogenolysis, and gluconeogenesis, which could be vital in starvation.

Fifth, in a setting where hyperglycemia is triggered by surgery or trauma, the equivalent of acute ischemia/reperfusion, the delay between onset of hyperglycemia and the start of glycemic control is short. In contrast, when ICU patients already suffered from chronic illness prior to ICU admission and hyperglycemia was present for a longer time, adaptive changes to protect the cells against elevated extracellular glucose may have been induced such that acute lowering of blood glucose may be harmful. Alternatively, the time window for prevention of toxicity may have passed and irreversible damage done.⁴⁸ Such a mechanism was suggested by the pooled analysis of the two Leuven trials⁴² and by the different results of RCTs on glucose control in patients with type 2 diabetes.⁴⁹^–⁵⁴

Finally, insulin therapy induces shift of potassium from the extracellular to the intracellular compartment. This may induce hypokalemia and hypokalemia-induced arrhythmias. By using arterial blood and an accurate point-of-care blood gas analyzer for glucose monitoring with each blood glucose check, potassium levels can be measured and corrected when needed.

All these differences may have contributed to the different outcomes in different studies. It has become clear that results from single-center, proof-of-concept studies cannot simply be repeated in large multicenter effectiveness trials, certainly when studying the effects of a complex intervention which is too often incompletely implemented in the repeat studies.⁵⁵ Hence, in reality, such studies did not investigate the same intervention as the proof-of-concept study.

Biological Rationale for Tight Glycemic Control

Research using human material, animal models, and in vitro systems has unraveled potential mechanistic explanations for the beneficial effects of TGC (Table 164-1). As in diabetes mellitus, insulinization to lower blood glucose levels exerts its effects on a wide array of biological pathways. Striving for metabolic control and inhibiting excess inflammation and mitochondrial damage seem of chief importance. Further molecular biology research will not only be essential to fine-tune TGC with other metabolic treatment strategies, it will also contribute to the quest to explain the potential harm of glucose lowering in critical illness.

TABLE 164-1 Studies of Biological Effects of Tight Glycemic Control Also Point to Its Potential Benefit

Implications for Daily Practice

The failure to repeat the results from well-controlled, meticulously executed, proof-of-concept studies in large pragmatic confirmation trials has indicated that the TGC is not yet ready to be broadly implemented in every ICU across the globe (Figure 164-3). This does not undermine the scientific validity of the benefits of TGC in critically ill patients. Blood glucose levels should be normalized as much as safely possible without causing a too-rapid lowering of blood glucose, without an increase in the incidence of hypoglycemia, and without large blood glucose fluctuations. Therefore it is advisable to gradually tighten glycemic control under diligent monitoring of the safety aspects. Nevertheless, three conditions should always be met:

1 Accurate and frequent blood glucose measurements as a reliable invasive, continuous glucose sensor is not yet available. Capillary blood samples are unreliable in the ICU and should never be used. Blood glucose measurements on on-site blood gas analyzers are currently the preferred devices. However, the use of a single handheld blood glucose meter with an acceptable error range and using arterial blood may be an alternative.

2 Continuous intravenous insulin administration using accurate syringe pumps

3 Thorough training of ICU healthcare providers (i.e., physicians and nurses) in the execution of the complex intervention of TGC. This stimulates intuitive and anticipating decision making, as computer algorithms to assist in TGC still have to show their benefit on patient outcome.