To evaluate the effect of using supplemental parenteral nutrition compared to early enteral nutrition alone on nutritional and clinical outcomes.
A multicenter, observational study.
Two hundred twenty-six intensive care units from 29 Countries.
Mechanically ventilated critically ill adult patients that remained in the intensive care unit for >72 hrs and received early enteral nutrition within 48 hrs from admission.
Data were collected on patient characteristics and daily nutrition practices for up to 12 days. Patient outcomes were recorded after 60 days.
Measurements and Main Results:
We compared the outcomes of patients who received early enteral nutrition alone, early enteral nutrition + early parenteral nutrition, and early enteral nutrition + late parenteral nutrition (after 48 hrs of admission). Cox regression analyses were conducted to determine the effect of feeding strategy, adjusted for other confounding variables, on time to being discharged alive from hospital. A total of 2,920 patients were included in this study; 2562 (87.7%) in the early enteral nutrition group, 188 (6.4%) in the early parenteral nutrition group, and 170 (5.8%) in the late parenteral nutrition group. Adequacy of calories and protein was highest in the early parenteral nutrition group (81.2% and 80.1%, respectively) and lowest in the early enteral nutrition group (63.4% and 59.3%) (p < .0001). The 60-day mortality rate was 27.8% in the early enteral nutrition group, 34.6% in the early parenteral nutrition group, and 35.3% in the late parenteral nutrition group (p = .02). The rate of patients discharged alive from hospital was slower in the group that received early parenteral nutrition (unadjusted hazard ratio 0.75, 95% confidence interval 0.59–0.96) and late parenteral nutrition (hazard ratio 0.64, 95% confidence interval 0.51–0.81) (p = .0003) compared to early enteral nutrition. These findings persisted after adjusting for known confounders.
The supplemental use of parenteral nutrition may improve provision of calories and protein but is not associated with any clinical benefit.