Skip to main content
Fig. 1 | BMC Geriatrics

Fig. 1

From: Overweight or obese BMI is associated with earlier, but not later survival after common acute illnesses

Fig. 1

Adjusted Hazard (and Odds) Ratios for Mortality in Primary and Sensitivity Analyses Panel a shows 1-year mortality results. Panel b shows 5-year mortality, conditional on survival to one year. The adjusted hazard ratio of overweight or obese BMI, relative to normal BMI (18.5 to < 25.0 kg/m2) is shown. The primary analysis is in white, while sensitivity analyses are shown in color: (1, red) excluding patients with unknown BMI trajectory or declining BMI trajectory prior to hospitalization; (2, orange) adjusting for acute illness severity; (3, yellow) logistic regression model; (4, green) include just one randomly selected hospitalization per person; (5, blue) analysis of age- and sex-matched pairs; and (6, purple) adjustment for pre-morbid disability and select co-morbidities. All models adjust for age, sex, race, marital status, education, smoking status, admission year, number of hospitalizations, household wealth, and household income.

Error bars represent the 95% CI for the hazard (or odds) ratio. A hazard or odds ratio of 1.0 indicates no association. A hazard ratio > 1.0 represents a positive association (overweight or obese BMI is associated with greater mortality), while a hazard ratio < 1.0 represents a negative association (overweight or obese BMI is associated with lower mortality)

Back to article page