Hyperglycemia Management

Poststroke hyperglycemia (PSH) affects more than 50% of people after acute ischemic stroke (AIS) and can be an acute stress reaction secondary to neuro-endocrine dysfunction or a result of underlying diabetes.^1-3 The definition of PSH varies widely in the literature with cutoffs from 109.8 mg/dL (6.1 mmol/L) to 180 mg/ dL (9.9 mmol/L).^3,4 There is ample evidence linking PSH with poor functional outcome and hemorrhagic transformation.^3,5-11 It remains unknown, however, whether PSH is the cause of poor outcomes or merely an epiphenomenon of other factors contributing to poor outcomes, such as large infarct size and underlying comorbidities. This review summarizes the evidence regarding the relationship between PSH and acute stroke outcomes, with emphasis on the role of in-hospital glycemic control after AIS.

Hyperglycemia and Outcomes After AIS

An association between PSH and poor outcomes has been reported repeatedly.^3,5-11 Post hoc analyses of clinical trials show associations between admission hyperglycemia and poor functional outcome at 3 months, neurologic deterioration, and hemorrhagic complications,⁷ but not with the risk of symptomatic intracranial hemorrhage.⁸

The relationship between PSH and functional and safety outcomes depends on multiple factors, including admission blood pressure (BP), underlying diabetes, and recanalization status. In a clinical trial of recombinant tissue plasminogen activator (rt-tPA), the relation between admission glucose and functional outcome depended on admission BP. As admission BP increased, the detrimental risk of hyperglycemia progressively increased. In a meta-analysis of 32 studies evaluating the relationship between admission blood glucose (BG) level and AIS outcomes, there was an increased risk of mortality (relative risk=3.28; 95% CI, 2.32 to 4.64) as well as poor functional recovery with a BG level of more than 110 to 126 mg/dL (6.1 to 7.0 mmol/L) in nondiabetic individuals only.³

Hyperglycemia and Reperfusion Therapies

Multiple large studies reported an association between PSH and poor outcome in people with AIS treated with intravenous (IV) tPA or endovascular thrombectomy.^4,5,7,12 A positive association between PSH with poor functional outcome at 3 months, mortality, and hemorrhagic complications has also been seen,⁵ but the association between admission hyperglycemia with mortality and symptomatic intracerebral hemorrhage was also statistically significant only in those without a history of diabetes. In a subsequent meta-analysis of 54 studies that evaluated the effect of diabetes or PSH on outcomes after thrombolysis, admission glucose level was inversely associated with favorable functional outcome (adjusted odds ratio [OR] 0.92; 95% CI, 0.90 to 0.94) and directly associated with symptomatic intracranial hemorrhage (adjusted OR 1.09; 95% CI, 0.104 to 1.14), with no significant difference in favorable outcome between those with a history of diabetes and those without (P-het=.54).¹²

Results have been similar among those treated with endovascular therapies. In a meta-analysis of 7 endovascular randomized trials, admission BG level was inversely associated with favorable functional outcome (adjusted OR, 0.93; 95% CI, 0.90 to 0.96) and associated with mortality (adjusted OR, 1.06; 95% CI, 1.03 to 1.09) and symptomatic intracranial hemorrhage (adjusted OR, 1.06; 95% CI, 1.02 to 1.11).⁴ BG level also modified endovascular therapy effects, with higher BG correlating with greater effects of endovascular therapy.

Hyperglycemia and Brain Injury

How PSH exacerbates brain injury during stroke is not yet fully understood. Acute hyperglycemia reduces intracellular antioxidant potential by 1) activating protein kinase C, thus increasing reactive oxygen species¹³; 2) depleting endothelium and vascular smooth muscle cells of nicotinamide adenine dinucleotide phosphate (NADPH)¹⁴; and 3) contributing to endothelial nitric oxide synthase (eNOS) dysfunction.¹⁴ Acute hyperglycemia also increases levels of plasminogen activator inhibitor 1 (PAI-1), promoting thrombus stabilization and microvascular plugging.¹⁴ These mechanisms contribute to the thrombotic, inflammatory, and vasoconstrictive effects of acute hyperglycemia. In addition, hyperglycemia can induce brain tissue lactic acidosis leading to cytotoxicity and eventual cell death, exacerbating the ischemic injury.^9,15

Animal studies using transient middle cerebral artery occlusion (tMCAO) demonstrated a significant reduction of cerebral blood flow during the reperfusion period and an increase in infarct size in hyperglycemic compared with normoglycemic rats.^16,17 Human imaging studies also report an association between hyperglycemia and larger infarcts. Serial diffusion weighted and perfusion weighted MRI, MR spectroscopy and BG measurements in people with ischemic stroke show glucose is associated with greater lactate production, and higher acute BG was positively associated with infarct growth.⁹

Clinical Trials of Hyperglycemia Management

Randomized clinical trials have not demonstrated benefit of intensive glycemic control after AIS (Table). A UK trial enrolled participants with acute stroke (ischemic or hemorrhagic) and a plasma glucose concentration of 108 to 306 mg/dL (6-17 mmol/L). Participants were randomly assigned to treatment with glucose-potassium-insulin (GKI) infusion with a target BG concentration 72 to 126 mg/dL (4-7 mmol/L) or normal saline infusion. At 3 months, there was no significant difference in death or disability. Rescue dextrose was administered for prolonged hypoglycemia in 73 participants in the treatment group; however, there was no significant difference in 90-day mortality in those who had rescue dextrose compared with those who did not and were also treated with GKI (P=0.55). Multiple limitations of this trial include the clinically small difference in the achieved BG level between groups (0.57 mmol/L or 10 mg/ dL), enrolling participants with different stroke types, and enrolling mostly nondiabetic participants (80%).¹⁸

Another larger randomized clinical trial compared an intensive glycemic control protocol (target BG 90-130 mg/ dL [4.4-7.2 mmol/L]) with a standard glycemic control protocol (target BG 80-179 mg/ dL [4.4-9.9 mmol/L]). Participants were adults with AIS and hyperglycemia (BG >110 mg/dL [6.1 mmol/L] if they had preexisting diabetes or ≥150 mg/dL [8.3 mmol/L]) if they did not have preexisting diabetes) within 12 hours of stroke onset. Despite a significant difference in the achieved mean BG levels between groups (118 mg/dL [6.6 mmol/L] vs 179 mg/dL (9.9 mmol/L) in the intensive and treatment groups, respectively), there was no benefit from the intensive treatment. Severe hypoglycemia (glucose level <40 mg/dL [2.2 mmol/L]) occurred in 2.6% of the intensive group.¹⁹

In addition, multiple small randomized preliminary trials and meta-analyses of glycemic control after AIS have been reported.^20,21 Overall, more intensive glycemic control protocols did not significantly improve clinical outcomes and increased occurrence of hypoglycemia.

Discussion and Treatment Recommendations

Despite a large body of evidence linking hyperglycemia with worse imaging and functional outcomes after AIS, randomized clinical trials have not shown a benefit from intensive glucose control. Potential explanations for this result are important to consider. Firstly, hyperglycemia during AIS can be a reflection of prestroke diabetes, and the associated cerebrovascular pathology is the primary determinant of outcomes. Controlling BG after stroke has occurred may be too late to affect the extent of brain injury during stroke. Thus, elevated BG level might be viewed as a marker rather than a cause of worse outcomes. Secondly, the treatment protocols used in clinical trials were initiated hours after stroke onset, perhaps too late to exert a beneficial effect, considering that time is brain during AIS. Thirdly, glucose levels achieved in the nonintensive treatment groups may have been sufficiently low to avoid poor outcomes (179 mg/ dL [9.9 mmol/dL] in the larger trial discussed). Testing individuals who had AIS with BG levels more than 250 mg/ dL (13.9 mmol/ dL) as a control group, for example, may have produced a different result. Fourthly, the duration of intensive glucose control in randomized trials may have been too short (72 hours in the larger trial). Perhaps longer intensive control would have enhanced stroke healing and recovery processes and shown improved outcomes. Additional considerations include achieved difference in BG levels between treatment groups. Although in the UK trial the difference was clinically small (10 mg/dL or 0.6 mmol/ dL), in the larger trial the difference was considerably larger (61 mg/dL or 3.4 mmol/ dL).

The 2019 American Heart Association (AHA)/American Stroke Association (ASA) acute stroke management guidelines recommend maintaining BG between 140 mg/dL (7.8 mmol/dL) and 180 mg/dL (10 mmol/dL) in the first 24 hours (moderate recommendations) after stroke.²² The European Stroke Organization (ESO) guidelines recommend against routine use of IV insulin to achieve tight glycemic control for improvement of functional outcome (weak recommendations), and do not specify a BG target range.²³ Considering the available evidence, it is reasonable to maintain BG between 140 and 180 mg/dL (7.8-10 mmol/dL) during hospitalization after AIS with SC or IV insulin when needed.

Because many acute stroke patients have difficulty swallowing or may be on a respirator, making daily carbohydrate intake temporarily uncertain, it is best to withhold all usual antidiabetic treatments and use periodic subcutaneous regular insulin injections, adjusted according to BG levels. To decrease BG fluctuations, rapid-acting insulins can be used immediately after meals based on the amount of carbohydrates consumed. For standard meals containing 60 g of carbohydrates, 4 units of rapid acting analog insulin is a reasonable starting dose. To further decrease BG fluctuations, long-acting insulins can be added when stable daily carbohydrate intake is established. Eventually, long-term antidiabetic treatment, including pills, can be safely resumed.