LAUSR

e64 www.ccmjournal.org January 2019 • Volume 47 • Number 1 The authors reply: We agree with Ji et al (1) that body weight loss is partly due to an increase of catabolism and/or an insufficient nutrition with an associated muscle loss. The database does not permit to answer this question since nutritional biomarkers are not recorded. The use of enteral but not parenteral nutrition was associated with the occurrence of bedsores in univariate analysis but not in multivariate analysis. However, the relationship between occurrence of bedsores and body weight variation (BWV) was not influenced by the nature of artificial nutrition (2). As far as the statistical modeling of the BWVs was concerned, we didn’t believe that a unique analysis considering the BWVs as a time-dependent covariate could be appropriate because such models could only estimate the global effect of the BWVs at time t. Yet, the effect of the BWVs is not linear over time. Consequently, we preferred our landmark analyses. As a result, we demonstrated that only a remaining overweight at day 5 and 6 was associated with an increased risk of death, but not at day 3 and 4, but also that only a weight loss at day 3 was significantly associated with the occurrence of bedsore. A sensitivity analysis using a cause-specific model instead of a subdistribution model gave similar results (3) (Table 1). Nutrition Support is a Potential Confounding Factor for the Relationship Between Body Weight Variation and the Risk of Bedsore