Temporal trends in the use of parenteral nutrition in critically ill patients

Background: Clinical practice guidelines recommend enteral over parenteral nutrition in critical illness and do not recommend early initiation. Few data are available on parenteral nutrition use or timing of initiation in the ICU or how this use may have changed over time. Methods: We used the Project IMPACT database to evaluate temporal trends in parenteral nutrition use (total and partial parenteral nutrition and lipid supplementation) and timing of initiation in adult ICU admissions from 2001 to 2008. We used x² tests and analysis of variance to examinecharacteristics of patients receiving parenteral nutrition and multilevel multivariate logistic regression models to assess parenteral nutrition use over time, in all patients and in specific subgroups. Results: Of 337,442 patients, 20,913 (6.2%) received parenteral nutrition. Adjusting for patient characteristics, the use of parenteral nutrition decreased modestly over time (adjusted probability, 7.2% in 2001-2002 vs 5.5% in 2007-2008, P <.001). Enteral nutrition use increased simultaneously (adjusted probability, 11.5% in 2001-2002 vs 15.3% in 2007-2008, P <.001). Use of parenteral nutrition declined most rapidly in emergent surgical patients, patients with moderate illness severity, patients in the surgical ICU, and patients admitted to an academic facility ( P ≤.01 for all interactions with year). When used, parenteral nutrition was initiated a median of 2 days (interquartile range, 1-3), after ICU admission and > 90% of patients had parenteral nutrition initiated within 7 days; timing of initiation of parenteral nutrition did not change from 2001 to 2008. Conclusions: Use of parenteral nutrition in US ICUs declined from 2001 through 2008 in all patients and in all examined subgroups, with the majority of parenteral nutrition initiated within the fi rst 7 days in ICU; enteral nutrition use coincidently increased over the same time period.