Is body mass index associated with mortality in patients with sepsis?

There is controversy whether patients with higher body mass indexes (BMI) have better survival after sepsis.

To determine if the association between BMI and mortality persists after adjustment for weight loss, comorbidities, and nutritional status.

In this single-center retrospective study of patients meeting Sepsis-3 criteria between July 10, 2009 and September 7, 2019, we used logistic regression to adjust BMI for weight loss, comorbidities, culture results, demographics, laboratory values, and processes of care. We transformed BMI using fractional polynomials. In a sensitivity analysis, we categorized BMI using the World Health Organization six levels of underweight, normal, overweight, class I, class II, and class III obesity.

5254 (25 %) of 20,990 patients died by 90 days. Unadjusted mortality was lowest at BMI of 32 kg/m². After adjustment, BMI showed only a minimal association with 90-day mortality. Compared to a patient with BMI = 25 kg/m², a patient with BMI = 15 kg/m² had a slightly increased odds ratio (1.022 with 95 % confidence interval (1.020, 1.031), while patients with BMI = 35 or 45 kg/m² had tiny reductions in risk, OR = 0.994 (0.992, 0.995) and 0.992 (0.988, 0.992), respectively. The sensitivity analysis using the World Health Organization levels had similar results. Compared to normal weight patients, underweight patients had a 1.004 (1.001, 1.007) odds ratio of dying, while Class II obese patients had a 0.994 (0.992, 0.996) of dying by 90-days.

After adjusting for comorbidities and nutritional status, BMI had only a tiny association with 90-day mortality.