Association between in-hospital weight change and 28-day mortality in adults with gram-negative rod bacteremia

Background: Body weight and fluid balance are critical parameters in the monitoring of patient status, as low body weight and fluid overload are associated with increased mortality in sepsis cases. While body weight reflects nutritional and fluid status, fluid balance is a major indicator of clinical stability. However, limited research has comprehensively evaluated the combined impact of in-hospital weight change and fluid balance on mortality in bacteremia patients, especially those undergoing treatment outside of ICUs.

Aim: This study aimed to investigate the association between in-hospital weight change, fluid balance, and 28-day mortality in patients with gram-negative rod (GNR) bacteremia, the objective being to determine the potential usefulness of these parameters in prediction of patient outcomes.

Methods: A retrospective cohort study on GNR bacteremia inpatients was conducted from April 2017 to March 2023 at St. Luke's International Hospital. Univariate analysis, using chi-square and t-tests, was followed by multivariate logistic regression analysis to explore the relationship between 28-day mortality and a number of variables, including in-hospital weight change, with significance level set at p < 0.05.

Results: Among 255 patients (230 survivors, 25 non-survivors), mean age was 76.3 years among survivors and 82.4 years among non-survivors (p = 0.05). In the survivor group, 49.1 % were female, compared to 44.0 % of the non-survivors (p = 0.78). In-hospital weight gain was significantly higher in non-survivors (3.81 %) compared to survivors (-0.81 %, p < 0.001). Non-survivors had higher positive daily fluid balance and lower daily total calorie intake (p < 0.05). After adjusting for other variables (age, sex, baseline clinical status and nutritional intake), multivariate analysis showed weight gain to be associated with 28-day mortality (odds ratio 1.10, 95 % CI: 1.01-1.20, p = 0.04).

Conclusion: Potentially influenced by fluid overload, in-hospital weight gain is associated with 28-day mortality. As a non-invasive and simple parameter, body weight shows potential as a practical tool for predicting outcomes, especially when accompanied by fluid balance assessments. Further prospective studies are needed to understand the multiple interactions between weight, fluid, and calorie intake.