In summary, for patients who received tight glycemic control:

Benefits in NNT

    No one was helped
    No one was helped

Harms in NNT

    7.6% higher risk of severe or symptomatic hypoglycemia
    1 in 13 were harmed (severe or symptomatic hypoglycemia)

View As:

Source: Cerecedo-Lopez CD, Cantu-Aldana A, Patel NJ, Aziz-Sultan MA, Frerichs KU, Du R. Insulin in the management of acute ischemic stroke: a systematic review and meta-analysis.World Neurosurg 2020;8750:e514–34

Study Population: A total of 2,734 patients with primarily acute ischemic stroke of differing severities and active hyperglycemia during in‐hospital treatment

Efficacy Endpoints: Mortality, independence, and functional outcome at 90 days

Harm Endpoints: Severe or symptomatic hypoglycemia

Narrative: Hyperglycemia in the setting of acute ischemic stroke is associated with increased mortality and worse neurologic outcome.1 Multiple mechanisms exist by which hyperglycemia may worsen ischemic brain injury, including increased inflammatory stress, calcium imbalance, accumulation of reactive oxygen species, and decreased blood perfusion, but it is unclear if treating hyperglycemia can reduce brain injury.2, 3, 4, 5 While earlier studies had smaller sample sizes,4, 6, 7 a 2019 large multicenter randomized controlled trial (RCT) was recently published,5 prompting the need for a new systematic review.

The systematic review and meta‐analysis summarized here included RCTs of patients experiencing acute ischemic stroke who had hyperglycemia at the time of or shortly after admission.8 Included RCTs utilized intravenous (IV) insulin for tight glycemic control (target glucose 70–135 mg/dL) as the intervention arm, used subcutaneous insulin therapy and/or placebo as the control arm, and evaluated at least one clinical outcome. Patients in the control arms received treatment with a wide range of glucose targets and regimens: these included numerical targets of < 300, 200, and 180 mg/dL. Others focused on mealtime glucose or sliding scales targeting 80 to 179 or < 150 mg/dL. There was no restriction on length of follow‐up. Outcomes included survival, functional outcome, and rates of severe or symptomatic hypoglycemia, defined as clinical and laboratory findings or laboratory findings only, based on the individual study definition. Each outcome was analyzed independently, and “independence” was defined as a favorable score on one of multiple scales. As a result, different scales were pooled together.

The review included 12 studies of 2,734 patients. The average age was 70, 50% of subjects were female, and 47% had known diabetes. Follow‐up ranged from 5 to 120 days. All studies used IV insulin infusion for the intervention, while control arms included standard therapy with insulin, carbohydrate‐restricted diet, or 0.9% normal saline placebo. Two studies accounted for the majority of the included patients.4, 5 Definition of hypoglycemia ranged from a glucose level of < 40 mg/dL at any one time to < 72 mg/dL for 30 minutes. Six or fewer RCTs examined clinical benefit (i.e., mortality, independence, mRS), and 11 RCTs examined clinical harm (i.e., hypoglycemia). Four of the larger studies, including the most recent study, had very low risk of bias. There was low heterogeneity (I2 ~ 0) between studies for neurological outcomes and modest heterogeneity (I2 = 28%) for hypoglycemia. There were no signs of publication bias. The authors found that tight glycemic control did not improve 90‐day survival (odds ratio [OR] = 1.0, 95% confidence interval [CI] = 0.8 to 1.2; six trials, n = 2,424), independence (OR = 0.95, 95% CI = 0.8 to 1.1; six trials, n = 2,424) or modified Rankin score. Tight glycemic control did, however, increase severe or symptomatic hypoglycemia (OR = 5.2, 95% CI = 1.7 to 15.9, absolute risk difference = 7.6%, number needed to harm = 13).8

Caveats: This meta‐analysis suggests tight glycemic control in patients with acute ischemic stroke increases severe or symptomatic hypoglycemia but does not improve survival or functional outcome. However, several limitations should be considered. Among the trials included in the review, one did not require imaging confirmation of ischemic stroke and included patients with intracerebral hemorrhage.4 Eight studies suffered from significant methodologic risk of bias, including unblinded outcome assessment. Included studies varied in their definition of hypoglycemia, and CIs were wide for the pooled outcome examining hypoglycemia. It is unknown whether a glucose level of < 72 mg/dL is clinically symptomatic or harmful. Studies also varied in terms of ischemic stroke severity included, in the scores utilized for the assessment of functional outcomes, and in control interventions. The most current trial from 2019, with 1,118 subjects, included patients with National Institutes of Health Stroke Scale of 3 to 22. Those at the extremes of ischemic stroke severity were therefore not included.5 Six trials were small and exploratory in nature, and follow‐up was short, offering no data for outcomes at ≥ 90 days. Also of note, two studies enrolled over 80% of subjects analyzed in the review, potentially limiting applicability to patients who meet the inclusion and exclusion criteria for those investigations. These two studies, comprising most of the data, were stopped early.4, 5 Early stoppage of trials tends to favor intervention arms and overestimate the effects.9

Other important considerations include heterogeneity in the control groups, including the specific treatments and serum glucose targets. Control group glucose targets and treatments varied: two trials used normal saline, one used carbohydrate restriction, and eight used insulin. This review therefore does not provide answers concerning optimal regimens, only a consensus that tight control (<135 mg/dL) with IV insulin provides no clinical benefit while substantially increasing episodes of hypoglycemia. Tight glucose control is also time‐ and resource‐intensive.

Despite these limitations, this analysis suggests tight glycemic control targeting serum glucose levels of 70 to 135 mg/dL with IV insulin therapy is associated with hypoglycemia and no discernible benefit. We have thus assigned a color recommendation of black (harm> benefits).

The original manuscript was published in Academic Emergency Medicine as part of the partnership between and AEM.

Author: Brit Long, MD; Alex Koyfman, MD; Michael Gottlieb, MD, RDMS
Supervising Editor: Shahriar Zehtabchi, MD

Published/Updated: June 5, 2020

  1. The Title Bar

    The title bar is color-coded with our overall recommendation.

    • Green: Benefits outweigh risks.
    • Yellow: Unclear risk/benefit profile.
    • Red: Benefits do not outweigh risks.
    • Black: Obvious harms, no clear benefits.
  2. Tip content...