Secondary Logo

Journal Logo

Institutional members access full text with Ovid®

Does Obesity Protect Against Death in Sepsis? A Retrospective Cohort Study of 55,038 Adult Patients*

Pepper, Dominique J. MD1; Demirkale, Cumhur Y. PhD1; Sun, Junfeng PhD1; Rhee, Chanu MD2,3; Fram, David PhD4; Eichacker, Peter MD1; Klompas, Michael MD2,3; Suffredini, Anthony F. MD1; Kadri, Sameer S. MD1

doi: 10.1097/CCM.0000000000003692
Feature Articles
Editor's Choice

Objectives: Observational studies suggest obesity is associated with sepsis survival, but these studies are small, fail to adjust for key confounders, measure body mass index at inconsistent time points, and/or use administrative data to define sepsis. To estimate the relationship between body mass index and sepsis mortality using detailed clinical data for case detection and risk adjustment.

Design: Retrospective cohort analysis of a large clinical data repository.

Setting: One-hundred thirty-nine hospitals in the United States.

Patients: Adult inpatients with sepsis meeting Sepsis-3 criteria.

Exposure: Body mass index in six categories: underweight (body mass index < 18.5 kg/m2), normal weight (body mass index = 18.5–24.9 kg/m2), overweight (body mass index = 25.0–29.9 kg/m2), obese class I (body mass index = 30.0–34.9 kg/m2), obese class II (body mass index = 35.0–39.9 kg/m2), and obese class III (body mass index ≥ 40 kg/m2).

Measurements: Multivariate logistic regression with generalized estimating equations to estimate the effect of body mass index category on short-term mortality (in-hospital death or discharge to hospice) adjusting for patient, infection, and hospital-level factors. Sensitivity analyses were conducted in subgroups of age, gender, Elixhauser comorbidity index, Sequential Organ Failure Assessment quartiles, bacteremic sepsis, and ICU admission.

Main Results: From 2009 to 2015, we identified 55,038 adults with sepsis and assessable body mass index measurements: 6% underweight, 33% normal weight, 28% overweight, and 33% obese. Crude mortality was inversely proportional to body mass index category: underweight (31%), normal weight (24%), overweight (19%), obese class I (16%), obese class II (16%), and obese class III (14%). Compared with normal weight, the adjusted odds ratio (95% CI) of mortality was 1.62 (1.50–1.74) for underweight, 0.73 (0.70–0.77) for overweight, 0.61 (0.57–0.66) for obese class I, 0.61 (0.55–0.67) for obese class II, and 0.65 (0.59–0.71) for obese class III. Results were consistent in sensitivity analyses.

Conclusions: In adults with clinically defined sepsis, we demonstrate lower short-term mortality in patients with higher body mass indices compared with those with normal body mass indices (both unadjusted and adjusted analyses) and higher short-term mortality in those with low body mass indices. Understanding how obesity improves survival in sepsis would inform prognostic and therapeutic strategies.

1Critical Care Medicine Department, Clinical Center, National Institutes of Health, Bethesda, MD.

2Department of Population Medicine, Harvard Medical School/Harvard Pilgrim Health Care Institute, Boston, MA.

3Department of Medicine, Brigham and Women’s Hospital, Boston, MA.

4Commonwealth Informatics, Waltham, MA.

*See also p. 735.

The findings and conclusions in this study are those of the authors and do not necessarily represent the official position of the National Institutes of Health.

Drs. Pepper, Demirkale, Sun, and Kadri had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis, including and especially any adverse effects. Drs. Pepper, Demirkale, Sun, Rhee, Fram, Eichacker, Klompas, Suffredini, and Kadri contributed substantially to the study design, data analysis, and interpretation. Drs. Pepper and Kadri drafted the article. Drs. Demirkale, Sun, Rhee, Fram, Eichacker, Klompas, and Suffredini revised it critically for important intellectual content. Drs. Pepper, Demirkale, Sun, Rhee, Fram, Eichacker, Klompas, Suffredini, and Kadri approve the final version to be published.

Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s website (

Supported, in part, by the National Institutes of Health Intramural Research Program, Clinical Center.

Dr. Pepper received other support from intramural funding from the National Institutes of Health (NIH). Drs. Pepper, Demirkale, Sun, Fram, Eichacker, Suffredini, and Kadri received support for article research from the NIH. Drs. Pepper, Demirkale, Sun, Suffredini, and Kadri disclosed government support. Dr. Rhee’s institution received support for article research from Agency for Healthcare Research and Quality (AHRQ) (grant number K08HS025008), and he received support for article research from AHRQ. Dr. Fram’s institution received funding from the NIH; he received funding from Commonwealth Informatics (stockholder); and he disclosed work for hire. Dr. Klompas’ institution received funding from the Centers for Disease Control and Prevention.

Address requests for reprints to: Dominique J. Pepper, MD, MBChB, Critical Care Medicine Department, National Institutes of Health, Building 10, Room 2C145 Bethesda, MD 20892. E-mail:

Copyright © 2019 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.