Admission hyperglycemia in critically ill sepsis patients : association with outcome and host response

Abstract

Objectives: To investigate whether admission hyperglycemia is associated with the presentation and/or outcome of sepsis, what the influence of hyperglycemia is on key host responses to sepsis, and whether hyperglycemia differentially affects patients with diabetes mellitus.

Design and setting: A substudy of a prospective observational cohort study was conducted in the intensive care of two tertiary hospitals between January 2011 and July 2013.

Patients: Of all consecutive critically ill sepsis patients, admission glucose was used to stratify patients in euglycemia (71-140 mg/dL), mild hyperglycemia (141-199 mg/dL), and severe hyperglycemia (≥ 200 mg/dL), and patients with hypoglycemia were excluded. Fifteen plasma biomarkers providing insight in key host responses implicated in sepsis pathogenesis were measured on admission.

Measurements and main results: Of 987 sepsis patients with admission glucose levels greater than 70 mg/dL, 519 (52.6%) had normal glucose levels, 267 (27.1%) had mild, and 201 (20.4%) severe hyperglycemia. Admission hyperglycemia was accompanied by mitigated alterations in plasma host response biomarker levels indicative of activation of the cytokine network, the vascular endothelium, and the coagulation system in patients without a history of diabetes. Severe, but not mild, admission hyperglycemia was associated with increased 30-day mortality (adjusted hazard ratio, 1.66 [95% CI, 1.24-2.23]), in both patients without diabetes (adjusted hazard ratio, 1.65 [95% CI, 1.12-2.42]) and with diabetes (adjusted hazard ratio, 1.91 [95% CI, 1.01-3.62]).

Conclusion: Admission hyperglycemia is associated with adverse outcome of sepsis irrespective of the presence or absence of preexisting diabetes by a mechanism unrelated to exaggerated inflammation or coagulation.