Hepatic Failure/GI/Endocrine Emergencies

Critically ill patients often do not have the “classic triad” of diabetes symptoms of polyuria,

polydipsia, and polyphagia.

Stress-induced hyperglycemia has been associated with illness severity.

Critically ill patients have increased release of “stress hormones” (e.g., cortisol and epinephrine)

and cytokines. These responses lead to both increased glucose production and insulin resistance,

which results in hyperglycemia.

Hyperglycemia is further exacerbated by infusions of dextrose-containing fluids, corticosteroids,

and exogenous sympathomimetic medication administration.

Management

Hyperglycemia was once considered a beneficial adaptive response in the critically ill population

and was not considered a treatment priority. In general, BG was only treated if it exceeded 200 mg/

dL.

Tighter glucose control garnered increased interest because hyperglycemia has independently been

associated with increased ICU mortality.

A treatment paradigm shift occurred in 2001 with the publication of the landmark “intensive

insulin therapy” study.

In this single-center study, surgical ICU patients who were receiving parenteral nutrition were

randomized to intensive intravenous insulin therapy (goal BG 80–110 mg/dL) or conventional

insulin therapy (goal BG 180–200 mg/dL).

ii.

Patients randomized to intensive insulin therapy had a significantly lower ICU mortality

rate than patients receiving conventional insulin therapy (4.6% vs. 8.0%, p<0.04). The ICU

mortality benefit was most pronounced in patients who stayed in the ICU (and received

intensive insulin therapy) for greater than 5 days (10.6% vs. 20.2%, p=0.005).

iii.

Patients randomized to intensive insulin therapy also less commonly developed bloodstream

infections, acute kidney injury, and ICU-acquired weakness (at the time termed critical-illness

polyneuropathy).

iv.

A study of medical ICU patients with an identical design from the same center was published

in 2006. The study detected no in-hospital mortality benefit with intensive insulin therapy

(37.3% vs. 40.0% with conventional insulin therapy, p=0.33). Patients in the intensive insulin

therapy arm less commonly developed acute kidney injury, had a shorter time to liberation

from mechanical ventilation, and had earlier discharge from the ICU and hospital. For patients

who stayed in the ICU for 3 days or more, in-hospital mortality was lower in the intensive

insulin therapy arm (43.0% vs. 52.5%, p=0.009), but the validity of this subgroup analysis has

been called into question because patients were not defined by a baseline characteristic.

In light of these findings, intensive insulin therapy was widely recommended by treatment

guidelines (including the 2008 Surviving Sepsis Campaign guidelines) and often implemented

into practice as a standard of care.

Because of concern with the single-center nature of the aforementioned studies, unblinded

design, and large relative mortality benefit, three multicenter trials were designed and

conducted.

ii.

One multicenter study was terminated early because of safety concerns. Patients randomized

to intensive insulin therapy more commonly developed severe hypoglycemia (BG of 40 mg/

dL or less) than those allocated to conventional insulin therapy (17.0% vs. 4.1%, p<0.001).

Intensive insulin therapy was not associated with a benefit in 28-day mortality (24.7% vs.

26.0%, p=0.74), but the study was inadequately powered to assess this outcome.