Ovid: Welcome to Ovid

Circulating high-mobility group box 1 (HMGB1) concentrations are elevated in both uncomplicated pneumonia and pneumonia with severe sepsis *.
Angus, Derek C. MD, MPH; Yang, LiHong PhD; Kong, Lan PhD; Kellum, John A. MD; Delude, Russell L. PhD; Tracey, Kevin J. MD; Weissfeld, Lisa PhD; for the GenIMS Investigators
[Article] Critical Care Medicine. 35(4):1061-1067, April 2007.

(Format: HTML, PDF)

Objective: High-mobility group box 1 (HMGB1) has been proposed as a late mediator of sepsis, but human data are sparse and conflicting. We describe plasma HMGB1 concentrations in humans with community-acquired pneumonia (CAP), the most common cause of severe sepsis, and test the hypotheses that HMGB1 levels are higher in CAP than healthy controls, higher in CAP with severe sepsis than CAP without severe sepsis, and higher in severe sepsis nonsurvivors than survivors.

Design: Random, outcome-stratified sample from a prospective study of 1,895 subjects hospitalized with CAP.

Setting: Twenty-eight U.S. teaching and community hospitals.

Patients: There were 122 CAP subjects (43 never developed severe sepsis, 49 developed severe sepsis and survived hospitalization, and 30 developed severe sepsis and died) and 38 healthy controls.

Interventions: None.

Measurements and Main Results: Median day of onset of severe sepsis was day of admission. HMGB1 was measured daily for the first week and analyzed using repeated-measures models with and without multivariable adjustment for baseline characteristics. HMGB1 concentrations were higher in CAP subjects compared with controls (median concentration on day of admission vs. controls, 190 vs. 0 ng/mL, p = .0001; 93.7% of all CAP measurements were elevated). HMGB1 remained elevated throughout the hospital course with no significant trend (p = .64) and did not differ between those with and without severe sepsis (p = .30). HMGB1 concentrations were higher in severe sepsis nonsurvivors than survivors (p = .001). HMGB1 concentrations remained elevated at discharge (median final HMGB1 measure, 176 ng/mL). Findings persisted in multivariable models and were robust to sensitivity analyses using alternative definitions of severe sepsis.

Conclusions: HMGB1 is elevated in almost all CAP subjects, and higher circulating HMGB1 is associated with mortality. But immunodetectable HMGB1 levels were also persistently elevated in those patients who fared well. Thus, additional work is needed to understand the biological activities of serum HMGB1 in sepsis.