Utility of Lean Body Mass Equations and Body Mass Index for Predicting Outcomes in Critically Ill Adults with Sepsis: A Retrospective Study

Abstract

Lean body mass is a significant component of survival from sepsis. Several equations can be used for calculating lean body mass based on age, sex, body weight, and height. We hypothesized that lean body mass is a better predictor of outcomes than the body mass index (BMI). This study used a multicenter cohort study database. The inclusion criteria were age ≥18 years and a diagnosis of sepsis or septic shock. BMI was classified into four categories: underweight (<18.5 kg/m²), normal (≥18.5-<25 kg/m²), overweight (≥25-<30 kg/m²), and obese (≥30 kg/m²). Four lean body mass equations were used and categorized on the basis of quartiles. The outcome was in-hospital mortality among different BMI and lean body mass groups. Among 85,558 patients, 3916 with sepsis were included in the analysis. Regarding BMI, in-hospital mortality was 36.9%, 29.8%, 26.7%, and 27.9% in patients who were underweight, normal weight, overweight, and obese, respectively (p < 0.01). High lean body mass did not show decreased mortality in all four equations. In critically ill patients with sepsis, BMI was a better predictor of in-hospital mortality than the lean body mass equation at intensive care unit (ICU) admission. To precisely predict in-hospital mortality, ICU-specific lean body mass equations are needed.

Keywords: body mass index; intensive care unit; lean body mass; mortality; sepsis.

Figure 1

Flow of data selection from the Japanese Intensive Care Patient Database. Out of the 85,558 registry entries, the analysis included 3,916 patients. ICU: intensive care unit.

Figure 2

Mortality based on BMI (kg/m²). Mortality was different in the four BMI groups (p < 0.01). Underweight patients had significantly higher mortality than normal or overweight in post hoc analysis. * p < 0.05 in post hoc Bonferroni tests.

Figure 3

Mortality based on four lean body mass equations. Four equations include the formula reported by Kulkarni et al. [22], Weijs et al. [23], Janmahasatian et al. [24], and Hume et al. [25] Groups were divided into lowest quartile, Q1, to highest quartile, Q4. * p < 0.05 in post hoc Bonferroni tests. Q: quartile.